oracle® endeca information discovery integrator® endeca information discovery integrator :...

Oracle® Endeca Information DiscoveryIntegrator

Integrator Server Guide

Version 3.0.0 • March 2013

Copyright and disclaimerCopyright © 2003, 2013, Oracle and/or its affiliates. All rights reserved.

Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks oftheir respective owners. UNIX is a registered trademark of The Open Group.

This software and related documentation are provided under a license agreement containing restrictions onuse and disclosure and are protected by intellectual property laws. Except as expressly permitted in yourlicense agreement or allowed by law, you may not use, copy, reproduce, translate, broadcast, modify, license,transmit, distribute, exhibit, perform, publish or display any part, in any form, or by any means. Reverseengineering, disassembly, or decompilation of this software, unless required by law for interoperability, isprohibited.

The information contained herein is subject to change without notice and is not warranted to be error-free. Ifyou find any errors, please report them to us in writing.

If this is software or related documentation that is delivered to the U.S. Government or anyone licensing it onbehalf of the U.S. Government, the following notice is applicable:

U.S. GOVERNMENT END USERS: Oracle programs, including any operating system, integrated software,any programs installed on the hardware, and/or documentation, delivered to U.S. Government end users are"commercial computer software" pursuant to the applicable Federal Acquisition Regulation and agency-specific supplemental regulations. As such, use, duplication, disclosure, modification, and adaptation of theprograms, including any operating system, integrated software, any programs installed on the hardware,and/or documentation, shall be subject to license terms and license restrictions applicable to the programs. Noother rights are granted to the U.S. Government.

This software or hardware is developed for general use in a variety of information management applications. Itis not developed or intended for use in any inherently dangerous applications, including applications that maycreate a risk of personal injury. If you use this software or hardware in dangerous applications, then you shallbe responsible to take all appropriate fail-safe, backup, redundancy, and other measures to ensure its safeuse. Oracle Corporation and its affiliates disclaim any liability for any damages caused by use of this softwareor hardware in dangerous applications.

This software or hardware and documentation may provide access to or information on content, products andservices from third parties. Oracle Corporation and its affiliates are not responsible for and expressly disclaimall warranties of any kind with respect to third-party content, products, and services. Oracle Corporation andits affiliates will not be responsible for any loss, costs, or damages incurred due to your access to or use ofthird-party content, products, or services.

Oracle® Endeca Information Discovery Integrator : Integrator Server Version 3.0.0 • March 2013Guide

Oracle Endeca Supplement to Clover ServerThis supplement provides specific information about support and limitations when using the Clover Server asthe Oracle Endeca Integrator Server.

Supported Containers

Oracle Endeca Integrator Server is only supported in the following containers:

• Apache Tomcat

• Oracle WebLogic

While the Clover Server allows installation to other containers, installation into these containers is notsupported for Oracle Endeca Integrator Server.

Oracle® Endeca Information Discovery Integrator : Integrator Server Version 3.0.0 • March 2013Guide

iii

Table of Contents1. What is CloverETL Server .......................................................................................................... 12. Installation ............................................................................................................................... 3

Evaluation Server .................................................................................................................. 3Enterprise Server .................................................................................................................. 4

Apache Tomcat ............................................................................................................ 5Jetty ........................................................................................................................... 8IBM Websphere .......................................................................................................... 10Glassfish / Sun Java System Application Server ................................................................ 13JBoss ........................................................................................................................ 15Oracle WebLogic Server .............................................................................................. 17Possible issues during installation .................................................................................. 19

Memory Settings ................................................................................................................. 23Upgrading Server to Newer Version ....................................................................................... 23

3. Server Side Job files - Sandboxes ............................................................................................... 25Referencing files from the ETL graph or Jobflow ..................................................................... 26Sandbox Content Security and Permissions .............................................................................. 26Sandbox Content ................................................................................................................. 27Job config properties ........................................................................................................... 32

4. Viewing Job Runs - Executions History ....................................................................................... 355. Users and Groups .................................................................................................................... 38

LDAP authentication ............................................................................................................ 38Web GUI section Users ........................................................................................................ 40Web GUI section Groups ..................................................................................................... 43

6. Scheduling ............................................................................................................................. 45Timetable Setting ................................................................................................................ 45Tasks ................................................................................................................................ 48

7. Graph Event Listeners .............................................................................................................. 55Graph Events ...................................................................................................................... 55Listener ............................................................................................................................. 56Tasks ................................................................................................................................ 56Use cases ........................................................................................................................... 60

8. Jobflow Event Listeners ............................................................................................................ 63Jobflow Events ................................................................................................................... 63Listener ............................................................................................................................. 64Tasks ................................................................................................................................ 64

9. JMS messages listeners ............................................................................................................. 65Optional Groovy code .......................................................................................................... 66Message data available for further processing ........................................................................... 66

10. Universal event listeners ......................................................................................................... 69Groovy code ....................................................................................................................... 69

11. Manual task execution ............................................................................................................ 7012. File event listeners ................................................................................................................. 71

Observed file ...................................................................................................................... 71File Events ......................................................................................................................... 72Check interval, Task and Use cases ........................................................................................ 73

13. WebDAV ............................................................................................................................. 74WebDAV clients ................................................................................................................. 74WebDAV authentication/authorization .................................................................................... 74

14. Simple HTTP API ................................................................................................................. 76Operation help .................................................................................................................... 76Operation graph_run ............................................................................................................ 77Operation graph_status ......................................................................................................... 77Operation graph_kill ............................................................................................................ 78Operation server_jobs .......................................................................................................... 79Operation sandbox_list ......................................................................................................... 79

CloverETL Server

iv

Operation sandbox_content ................................................................................................... 79Operation executions_history ................................................................................................. 79Operation suspend ............................................................................................................... 81Operation resume ................................................................................................................ 81Operation sandbox_create ..................................................................................................... 82Operation sandbox_add_location ............................................................................................ 82Operation sandbox_remove_location ....................................................................................... 82Cluster status ...................................................................................................................... 83

15. JMX mBean .......................................................................................................................... 84JMX configuration .............................................................................................................. 84Operations .......................................................................................................................... 87

16. SOAP WebService API ........................................................................................................... 88SOAP WS Client ................................................................................................................ 88SOAP WS API authentication/authorization ............................................................................. 88

17. Launch Service ...................................................................................................................... 89Launch Service Overview ..................................................................................................... 89Deploying Graph in Launch Service ....................................................................................... 89Designing the ETL graph/Jobflow for Launch Service ............................................................... 90Configuring the job in CloverETL Server web GUI ................................................................... 90Sending the Data to Launch Service ....................................................................................... 94Results of the Graph Execution ............................................................................................. 94

18. Configuration ........................................................................................................................ 96Config Sources and Their Priorities ........................................................................................ 96Examples of DB Connection Configuration .............................................................................. 97

Embedded Apache Derby ............................................................................................. 97MySQL ..................................................................................................................... 98DB2 .......................................................................................................................... 98Oracle ...................................................................................................................... 100MS SQL .................................................................................................................. 101Postgre SQL ............................................................................................................. 101JNDI DB DataSource ................................................................................................. 101

List of Properties ............................................................................................................... 10219. Graph parameters ................................................................................................................. 106

Another sets of parameters according the type of execution ....................................................... 106executed from Web GUI ............................................................................................. 106executed by Launch Service invocation ......................................................................... 106executed by HTTP API run graph operation invocation .................................................... 106executed by RunGraph component ................................................................................ 106executed by WS API method executeGraph invocation ..................................................... 107executed by task "graph execution" by scheduler ............................................................. 107executed from JMS listener ......................................................................................... 107executed by task "graph execution" by graph event listener ............................................... 107executed by task "graph execution" by file event listener .................................................. 108

How to add another graph parameters ................................................................................... 108Additional "Graph Config Parameters" .......................................................................... 108Task "execute_graph" parameters ................................................................................. 108

20. Recommendations for transformations developers ...................................................................... 109Add external libraries to app-server classpath ......................................................................... 109Another graphs executed by RunGraph component may be executed only in the same JVM instance........................................................................................................................................ 109

21. Logging .............................................................................................................................. 110Main logs ......................................................................................................................... 110Graph run logs .................................................................................................................. 110

22. Extensibility (Embedded OSGi framework) ............................................................................... 111Groovy Code API .............................................................................................................. 111Embedded OSGi framework ................................................................................................ 112

23. Extensibility CloverETL engine plugins ................................................................................... 11424. Clustering ........................................................................................................................... 115

CloverETL Server

v

High Availability ............................................................................................................... 115Scalability ........................................................................................................................ 115

Transformation Requests ............................................................................................. 116Parallel Data Processing ............................................................................................. 116

Recommendations for Cluster Deployment ............................................................................. 120Example of Distributed Execution ........................................................................................ 120

Details of the Example Transformation Design ............................................................... 121Scalability of the Example Transformation ..................................................................... 123

Cluster configuration .......................................................................................................... 125Mandatory properties .................................................................................................. 125Optional properties .................................................................................................... 126Example of 2 node cluster configuration ........................................................................ 126Load balancing properties ........................................................................................... 127

25. Temp Space Management ...................................................................................................... 129Overview ......................................................................................................................... 129Setup ............................................................................................................................... 129

Adding Temp Space ................................................................................................... 130

1

Chapter 1. What is CloverETL ServerCloverETL Server (CS) is the integrating member of CloverETL products family. It introduces the powerful Clovertool into the world of corporate applications. CloverETL Server itself is an enterprise class application, thus it isshipped as WAR file (WAR stands for Web Archive). CS is tested and works on a range of application servers:Apache Tomcat, Jetty, IBM Websphere, Sun Glassfish, JBoss or Oracle Weblogic. Basically, CS is a runtimeenvironment for graphs, which brings new possibilities of integrating Clover with your own software. WhereasClover Engine can be integrated only as an embedded library, CS implements several interfaces which can becalled by other applications using common protocols like HTTP. In addition, CS implements some thread andmemory management optimizations.

Chapter 1. What isCloverETL Server

2

Table 1.1. CloverETL Server and CloverETL Engine comparison

CloverETL Server CloverEngine as executable tool

possibilities of executinggraphs

by calling http (or JMX, etc.) APIs (Seedetails in Chapter 14, Simple HTTPAPI (p. 76).)

by executing external process or bycalling java API

engine initialization during server startup init is called for each graph execution

thread and memoryoptimalization

threads recycling, graphs cache, etc. not implemented

scheduling scheduling by timetable, onetime trigger,logging included

external tools (i.e. Cron) can bes used

statistics each graph execution has its own log fileand result status is stored; each eventtriggered by the CS is logged

not implemented

monitoring If graph fails, event listener will benotified. It may send email, executeshell command or execute another graph.See details in Chapter 7, Graph EventListeners (p. 55). Additionally serverimplements various APIs (HTTP andJMX) which may be used for monitoringof server/graphs status.

JMX mBean can be used while graph isrunning

storage of graphs andrelated files

graphs are stored on server file system inso called sandboxes

security and authorizationsupport

CS supports users/groups management,so each sandbox may have its ownaccess privileges set. All interfacesrequire authentication. See details inChapter 3, Server Side Job files -Sandboxes (p. 25).

passwords entered by user may beencrypted

integration capabilities CS provides APIs which can be calledusing common protocols like HTTP. Seedetails in Chapter 14, Simple HTTPAPI (p. 76).

CloverEngine library can be used asembedded library in client's Java code or itmay be executed as separated OS processfor each graph.

development of graphs CS supports team cooperation above oneproject (sandbox). CloverETL Designerwill be integrated with CS in furtherversions.

scalability CS implements horisontal scalabilityof transformation requests as well asdata scalability. See details in Chapter24, Clustering (p. 115) In additionCloverEngine implements is verticalscalability nativelly.

Clover Engine implements vertivalscalability

jobflow CS implements various jobflowcomponents. See details in theCloverETL manual.

Clover Engine itself has limited support ofjobflow.

3

Chapter 2. InstallationFollowing sections describe two different installation types. the section called “Evaluation Server” (p. 3) forquick and most simple installation without configuration and the section called “Enterprise Server” (p. 4) forfurther testing and production on choosen app-container and database.

Evaluation Server

The default installation of CloverETL Server does not need any extra database server. It uses the embeddedApache Derby DB. What is more, it does not need any subsequent configuration. CloverETL Server configuresitself during the first startup. Database tables and some necessary records are automatically created on first startupwith an empty database. In the Sandboxes section of the web GUI, you can then check that sandboxes and a fewdemo graphs are created there.

If you need to evaluate CloverETl Server features which need any configuration changes, e.g. sending emails,LDAP authentication, clustering, etc., or the CloverETL Server must be evaluated on another application containerthen Tomcat, please proceed with the common installation: the section called “Enterprise Server” (p. 4)

Installation of Apache Tomcat

CloverETL Server requires Apache Tomcat version 6.0.x to run.

If you have Apache tomcat already installed, you can skip this section.

1. Download the ZIP with binary distribution from http://tomcat.apache.org/download-60.cgi. Tomcat may beinstalled as a service on Windows OS as well, however there may be some issues with access to file system,so it's not recommended aproach for evaluation.

2. After you download the zip file, unpack it.

3. Run Tomcat by [tomcat_home]/bin/startup.sh (or [tomcat_home]/bin/startup.bat onWindows OS).

4. Check whether Tomcat is running on URL: http://localhost:8080/. Apache Tomcat info page should appear.

5. Apache Tomcat is installed.

If in need of detailed installation instructions, go to: http://tomcat.apache.org/tomcat-6.0-doc/setup.html

Installation of CloverETL Server

1. Check if you meet prerequisites:

• JDK or JRE v. 1.6.x or higher

• JAVA_HOME or JRE_HOME environment variable has to be set.

• Apache Tomcat 6.0.x is installed. See Installation of Apache Tomcat (p. 3) for details.

2. Set memory limits and other switches. See section the section called “Memory Settings” (p. 23) for details.

Create setenv file:

Unix-like systems: [tomcat]/bin/setenv.sh

http://tomcat.apache.org/download-60.cgi

http://localhost:8080/

http://tomcat.apache.org/tomcat-6.0-doc/setup.html

Chapter 2. Installation

4

export CATALINA_OPTS="$CATALINA_OPTS -XX:MaxPermSize=512m -Xms128m -Xmx2048m" export CATALINA_OPTS="$CATALINA_OPTS -Dderby.system.home=$CATALINA_HOME/temp -server"echo "Using CATALINA_OPTS: $CATALINA_OPTS"

Windows systems: [tomcat]/bin/setenv.bat

set CATALINA_OPTS="%CATALINA_OPTS% -XX:MaxPermSize=512m -Xms128m -Xmx2048m" set CATALINA_OPTS="%CATALINA_OPTS% -Dderby.system.home=%CATALINA_HOME%/temp -server"echo "Using CATALINA_OPTS: %CATALINA_OPTS%"

3. Download the web archive file (clover.war) containing CloverETL Server for Apache Tomcat andclover-license.war containing valid license.

4. Deploy both WAR files: clover.war and clover-license.war to [tomcat_home]/webappsdirectory.

To avoid deployment problems, Tomcat should be down during the copying.


6. Check whether CloverETL Server is running on URLs:

• Web-app root

http://[host]:[port]/[contextPath]

The default Tomcat port for the http connector is 8080 and the default contextPath for CloverETL Serveris "clover", thus the default URL is:

http://localhost:8080/clover/

• Web GUI

http://[host]:[port]/[contextPath]/gui http://localhost:8080/clover/gui

Use default administrator credentials to access the web GUI: username - "clover", password - "clover".

7. CloverETL Server is now installed and prepared for basic evaluation. There are couple of sandboxes withvarious demo transformations installed.

Enterprise Server

This section describes instalation of CloverETL Server on various app-containers in detail, also describes the wayshow to configure the server. If you need just quickly evaluate CloverETl Server features which don't need anyconfiguration, evaluation installation may be suitable: the section called “Evaluation Server” (p. 3)

CloverETL Server for enterprise environment is shipped as a Web application archive (WAR file). Thus standardmethods for deploying a web application on you application server may be used. However each application serverhas specific behavior and features. Detailed information about their installation and configuration can be foundin the chapters below.

List of suitable containers:

• Apache Tomcat (p. 5)

• Jetty (p. 8)

• IBM Websphere (p. 10)


http://localhost:8080/clover/gui


5

• Glassfish / Sun Java System Application Server (p. 13)

• JBoss (p. 15)

• Oracle WebLogic Server (p. 17)

In case of problems during your installation see Possible issues during installation (p. 19).

Apache Tomcat

Installation of Apache Tomcat

CloverETL Server requires Apache Tomcat version 6.0.x to run.

If you have Apache tomcat already installed, you can skip this section.

1. Download the binary distribution from http://tomcat.apache.org/download-60.cgi.

2. After you download the zip file, unpack it.


4. Check whether Tomcat is running on URL: http://localhost:8080/. Apache Tomcat info page should appear.

5. Apache Tomcat is installed.

If in need of detailed installation instructions, go to: http://tomcat.apache.org/tomcat-6.0-doc/setup.html


1. Download the web archive file (clover.war) containing CloverETL Server for Apache Tomcat.


• JDK or JRE v. 1.6.x or higher

• JAVA_HOME or JRE_HOME environment variable has to be set.

• Apache Tomcat 6.0.x is installed. CloverETL Server is developed and tested with the Apache Tomcat 6.0.xcontainer (it may work unpredictably with other versions). See Installation of Apache Tomcat (p. 5)for details.

• It is strongly recommended you change default limits for the heap and perm gen memory spaces.

See section the section called “Memory Settings” (p. 23) for details.

You can set the minimum and maximum memory heap size by adjusting the "Xms" and "Xmx" JVMparameters. You can set JVM parameters for Tomcat by setting the environment variable JAVA_OPTS inthe [TOMCAT_HOME]/bin/setenv.sh file (if it does not exist, you may create it).

Create setenv file:

Unix-like systems: [tomcat]/bin/setenv.sh

http://tomcat.apache.org/download-60.cgi


http://tomcat.apache.org/tomcat-6.0-doc/setup.html


6

export CATALINA_OPTS="$CATALINA_OPTS -XX:MaxPermSize=512m -Xms128m -Xmx1024m" export CATALINA_OPTS="$CATALINA_OPTS -Dderby.system.home=$CATALINA_HOME/temp -server"echo "Using CATALINA_OPTS: $CATALINA_OPTS"

Windows systems: [tomcat]/bin/setenv.bat

set CATALINA_OPTS="%CATALINA_OPTS% -XX:MaxPermSize=512m -Xms128m -Xmx1024m" set CATALINA_OPTS="%CATALINA_OPTS% -Dderby.system.home=%CATALINA_HOME%/temp -server"echo "Using CATALINA_OPTS: %CATALINA_OPTS%"

As visible in the settings above, there is also switch -server. For performance reasons, it is recommendedto run the container in the "server" mode.

3. Copy clover.war (which is built for Tomcat) to [tomcat_home]/webapps directory.

Please note, that copying is not an atomic operation. If Tomcat is running, mind duration of the copying process!Too long copying might cause failure during deployment as Tomcat tries to deploy an incomplete file. Instead,manipulate the file when the Tomcat is not running.

4. War file should be detected and deployed automatically without restarting Tomcat.

5. Check whether CloverETL Server is running on URLs:

• Web-app root

http://[host]:[port]/[contextPath]



• Web GUI

http://[host]:[port]/[contextPath]/gui



Use default administrator credentials to access the web GUI: user name - "clover", password - "clover".

Configuration of CloverETL Server on Apache Tomcat

Default installation (without any configuration) is recommended only for evaluation purposes. For production, atleast DB data source and SMTP server configuration is recommended.

There are more ways how to set config properties.

Context Parameters (Available on Apache Tomcat)

Some application servers allow setting context parameters without modifying the WAR file. This way ofconfiguration is recommended for Tomcat.

On Tomcat, it is possible to specify context parameters in the context configuration file - [tomcat_home]/conf/Catalina/localhost/clover.xml. The file is created automatically just after deploying theCloverETL Server web application.




7

To specify a property, add this element:

<Parameter name="[propertyName]" value="[propertyValue]" override="false" />

To modify Tomcat context params, add this to the context config file (modify credentials accordingly):

<Parameter name="jdbc.driverClassName" value="..." override="false" /><Parameter name="jdbc.url" value="..." /><Parameter name="jdbc.username" value="..." override="false" /><Parameter name="jdbc.password" value="..." override="false" /><Parameter name="jdbc.dialect" value="..." override="false" />

Note

Special characters you type in the context file have to be specified as XML entities. For instance,ampsersand "&" as "&" etc.

Properties File on Specified Location

Example of such a file:

jdbc.driverClassName=...jdbc.url=...jdbc.username=...jdbc.password=...jdbc.dialect=...

Which location the common properties file is loaded from is specified by the system property or environmentvariable clover_config_file (clover.config.file). This is a recommended way of configuring ifcontext parameters cannot be set in application server.

On Apache Tomcat, you can set the system property in the [TOMCAT_HOME]/bin/setenv.sh file (if it doesnot exist, you may create it). Just add: JAVA_OPTS="$JAVA_OPTS -Dclover_config_file=/path/to/cloverServer.properties".

Installation of CloverETL Server License

To be able to execute graphs, CloverETL Server requires a valid license. You can install CloverETL Server withoutany license, but no graph will be executed.

There are two ways of installing license on Tomcat. A simpler way is a separate web application clover-license.war. However, in cluster environment, configuring the plain license file has to be done (common forall application containers).

a) Separate License WAR

1. Download the web archive file clover-license.war.

2. Copy clover-license.war to the [tomcat_home]/webapps directory.

3. The war file should be detected and deployed automatically without restarting Tomcat.

4. Check whether the license web-app is running on:


8

http://[host]:[port]/clover-license/ (Note: contextPath clover-license is mandatoryand cannot by changed)

b) License.file Property

Alternatively, configure the server's "license.file" property. Set its value to full path to the license.dat file.

Note

CloverETL license can be changed any time by re-deploying clover-license.war.Afterwards, you have to let CloverETL Server know the license has changed.

• Go to server web GUI →Monitoring →License

• Click Reload license.

• Alternatively, you can restart the CloverETL Server application.

Warning: Keep in mind that during the WAR file redeployment, directory [tomcat_home]/webapps/[contextPath] has to be deleted. If Tomcat is running, it should do it automatically.Still, we suggest you check it manually, otherwise changes will not take any effect.

Apache Tomcat on IBM AS/400 (iSeries)

To run CloverETL Server on the iSeries platform, there are some additional settings:

1. Declare you are using Java 6.0 32-bit

2. Run java with parameter -Djava.awt.headless=true

To configure this you can modify/create a file [tomcat_home]/bin/setenv.sh which contains:

JAVA_HOME=/QOpenSys/QIBM/ProdData/JavaVM/jdk50/32bit

JAVA_OPTS="$JAVA_OPTS -Djava.awt.headless=true"

Jetty


1. Download the web archive file (clover.war) containing the CloverETL Server application which is builtfor Jetty.

2. Check if prerequisites are met:

• JDK or JRE version 1.6.x or higher

• Jetty 6.1.x - only this particular version is supported

All jetty-6 releases are available from http://jetty.codehaus.org/jetty/. Jetty 7 is not supported (as of Jetty 7,there have been huge differences in distribution packages as it is hosted by the Eclipse Foundation.

http://jetty.codehaus.org/jetty/


9

• Memory allocation settings

It involves JVM parameters: -Xms -Xmx (heap memory) and -XX:MaxPermSize (classloaders memorylimit). See section the section called “Memory Settings” (p. 23) for details.

You can set the parameters by adding

JAVA_OPTIONS='$JAVA_OPTIONS -Xms128m -Xmx1024m -XX:MaxPermSize=256m'

to [JETTY_HOME]/bin/jetty.sh

3. Copy clover.war to [JETTY_HOME]/webapps.

4. Create a context file clover.xml in [JETTY_HOME]/contexts and fill it with the following lines:

<?xml version="1.0" encoding="ISO-8859-1"?><!DOCTYPE Configure PUBLIC "-//Jetty//Configure//EN" "http://www.eclipse.org/jetty/configure.dtd"><Configure class="org.mortbay.jetty.webapp.WebAppContext"> <Set name="contextPath">/clover</Set> <Set name="war"><SystemProperty name="jetty.home" default="."/>/webapps/clover.war</Set></Configure>

clover.xml will be detected by Jetty and the application will be loaded automatically.

5. Run [JETTY_HOME]/bin/jetty.sh start (or [JETTY_HOME]/bin/Jetty-Service.exe onWindows OS).

Finally, you can check if the server is running e.g. on http://localhost:8080/test/.

Configuration of CloverETL Server on Jetty

Default installation (without any configuration) is recommended only for evaluation purposes. For production, atleast DB data source and SMTP server configuration is recommended.

There are more ways how to set config properties, yet the most common one is properties file in a specified location.

Properties file in Specified Location


jdbc.driverClassName=...jdbc.url=...jdbc.username=...jdbc.password=...jdbc.dialect=...license.file=/path/to/license.dat

The common properties file is loaded from a location which is specified by the environment/system propertyclover_config_file or clover.config.file. This is a recommended way of configuring Jetty.

On Jetty, you can set system property in the [JETTY_HOME]/bin/jetty.sh file. Add:

http://localhost:8080/test/


10

JAVA_OPTIONS="$JAVA_OPTIONS -Dclover_config_file=/path/to/cloverServer.properties"

Installation of CloverETL Server license

In order to execute graphs, CloverETL Server requires a valid license file. Despite that, you can install CloverETLServer without a license, but no graph will be executed.

1. Get the license.dat file.

2. Set the CloverETL Server license.file parameter to the path to license.dat.

• Add "license.file" property to the config properties file (as decribed in Configuration of CloverETL Serveron Jetty (p. 9). Set its value to full path to the license.dat file.

• Restart Jetty.

There are more ways how to configure the license. See Chapter 18, Configuration (p. 96) for a descriptionof all possibilities.

Note

CloverETL license can be changed any time by replacing the license.dat file. Afterwards, youhave to let CloverETL Server know the license has changed.

• Go to server web GUI →Monitoring →License



IBM Websphere


1. Get the web archive file (clover.war) with CloverETL Server application which is built for Websphere.



• IBM Websphere 7.0 (see http://www.ibm.com/developerworks/downloads/ws/was/)


It involves JVM parameters: -Xms -Xmx and -XX:MaxPermSize. See section the section called “MemorySettings” (p. 23) for details.

You can set heap size and perm space in IBM Websphere's Integrated Solutions Console (by defaultaccessible at: http://[host]:10003//ibm/console/)

• Go to Servers →Application servers →[server1] (or another name of your server) →Process

Management →Java Virtual Machine

• There is the Maximum heap size field. Default value is only 256 MB, which is not enough for ETLtransformations.

http://www.ibm.com/developerworks/downloads/ws/was/


11

Figure 2.1. Adjusting Maximum heap size limit

• On the same page, there is Generic JVM arguments. Add the perm space limit there, e.g. like this:

-XX:MaxPermSize=512M

• Restart the server to confirm these changes.

3. Deploy WAR file

• Go to Integrated Solutions Console

( http://localhost:9060/ibm/console/)

• Go to Applications →New Application →New Enterprise Application

4. Configure logging

Websphere loggers do not use log4j by default. This may cause CloverETL Server logging to be ill-configured.As a result, some CloverETL Engine messages are missing in graph execution logs. Thus it is recommendedto configure Websphere properly to use log4j.

• Add a config file to the Websphere directory: AppServer/profiles/AppSrv01/properties/commons-logging.properties

• Insert these lines into the file:

http://localhost:9060/ibm/console/


12

priority=1 org.apache.commons.logging.LogFactory=org.apache.commons.logging.impl.LogFactoryImpl org.apache.commons.logging.Log=org.apache.commons.logging.impl.Log4JLogger

• Copy jar files from the clover.war/WEB-INF/lib archive to the AppServer/lib directory. Copyall files like commons-logging-*.jar and log4j-*.jar.

5. Try if the server is running

Provided you set clover.war as the application running with "clover" context path. Notice the port numberhas changed:

http://localhost:9080/clover

Configuration of CloverETL Server on IBM Websphere

Default installation (without any configuration) is recommended only for evaluation purposes. For production,configuring at least the DB data source and SMTP server is recommended.

There are more ways how to set config properties. The most common one is properties file in a specified location.

Properties File in Specified Location



Set system property (or environment variable) clover_config_file pointing to the config properties file.

• go to Integrated Solutions Console

(http://localhost:9060/ibm/console/)

• Go to Servers →Application servers →[server-name] →Java and Process Management →Process

Definition →Environment Entries

• Create system property named clover_config_file whose value is full path to config file named e.g.cloverServer.properties on your file system.

• This change requires restarting IBM Websphere.

Installation of CloverETL Server license

CloverETL Server requires a valid license for executing graphs. You can install CloverETL Server without alicense, but no graph will be executed.


2. Set CloverETL Server's license.file parameter to path to the license.dat file.

• Add "license.file" property to the config properties file as decribed in Configuration of CloverETL Serveron IBM Websphere (p. 12). Value of the property has to be full path to the license.dat file.

http://localhost:9080/clover

http://localhost:9060/ibm/console/


13

• Restart CloverETL Server.

There are other ways how to do this. The most direct one is to set system property or environment variableclover_license_file. (See Chapter 18, Configuration (p. 96) for description of all possibilities).

Note

Properly configured CloverETL license can be changed any time by replacing file license.dat.Then you have to let CloverETL Server know the license has changed.

• Go to web GUI →monitoring section →license tab

• Click button Reload license.


Glassfish / Sun Java System Application Server


1. Get CloverETL Server web archive file (clover.war) which is built for Glassfish (Tomcat).

2. Check if you meet prerequisites


• Glassfish (CloverETL Server is tested with V2.1)


It involves JVM parameters: -Xms -Xmx and -XX:MaxPermSize See section the section called“Memory Settings” (p. 23) for details.

You can set heap size and perm space in XML file [glassfish]/domains/domain1/config/domain.xml Add these sub-elements to <java-config>:

<jvm-options>-XX:MaxPermSize=512m</jvm-options><jvm-options>-Xmx2048m</jvm-options>

These changes require restarting Glassfish.

3. Deploy WAR file

• Copy WAR file to the server filesystem. CloverETL Server is packed in a WAR file of 100 MB approx., soit cannot be uploaded directly from your local filesystem using the Admin Console.

• Fill in attributes Application name and Context Root with "clover". Fill in path to the WAR file on theserver filesystem.

• Go to Glassfish Admin Console

It is accessible at http://localhost:4848/ by default; default username/password is admin/adminadmin

• Go to Applications →Web Applications and click Deploy.

• Submit form


14

Configuration of CloverETL Server on Glassfish

Default installation (without any configuration) is recommended only for evaluation purposes. For production,configuring at least the DB data source and SMTP server is recommended.





Set system property (or environment variable) clover_config_file pointing to the config properties file:

• Go to Glassfish Admin Console

By default accessible at http://localhost:4848/ with username/password: admin/adminadmin

• Go to Configuration →System Properties

• Create system property named clover_config_file whose value is full path to a file on your file systemnamed e.g. cloverServer.properties.

• This change requires restarting Glassfish.



License configuration is quite similar to WebSphere (p. 10).


2. Set CloverETL Server's license.file parameter to path to license.dat.

• Add "license.file" property to the config properties file as decribed in Properties File in SpecifiedLocation (p. 14). Set its value to full path to the license.dat file.

• Restart CloverETL Server.

There are of course other ways how to do this. The most direct one is setting system property or environmentvariable clover_license_file. (See Chapter 18, Configuration (p. 96) for description of allpossibilities).

Note

Properly configured CloverETL license can be changed any time by replacing license.dat.Next, you need to let CloverETL Server know the license has changed.

• Go to web GUI →Monitoring →License



15

• Alternatively, you can restart CloverETL Server.

JBoss


1. Get CloverETL Server web archive file (clover.war) which is built for JBoss.



• JBoss 6.0 or JBoss 5.1 - see http://www.jboss.org/jbossas/downloads

• Memory settings for jboss java process. See section the section called “Memory Settings” (p. 23) fordetails.

You can set the memory limits in [jboss-home]/bin/run.conf (run.conf.bat on Windows OS):

JAVA_OPTS="$JAVA_OPTS -XX:MaxPermSize=512m -Xms128m -Xmx2048m"

On Windows, perform steps analogic to the ones above.

3. Configure DB data source

Since JBoss does not work with embedded derby DB, a DB connection always has to be configured. We usedMySQL in this case

• Create datasource config file [jboss-home]/server/default/deploy/mysql-ds.xml

<datasources> <local-tx-datasource> <jndi-name>CloverETLServerDS</jndi-name> <connection-url>jdbc:mysql://localhost:3306/cloverServerDB</connection-url> <driver-class>com.mysql.jdbc.Driver</driver-class> <user-name>root</user-name> <password></password> </local-tx-datasource> </datasources>

Note

Special characters in the XML file have to be typed in as XML entities. For instance,ampsersand "&" as "&" etc.

JNDI name has to be exactly "CloverETLServerDS". The thing to do here is to set DB connectionparameters (connection-url, driver-class,user-name and password) to the created database.The database has to be empty before the first execution, the server creates its tables itself.

JNDI data source is the only way of configuring CloverETL Server DB connection in JBoss.

• Put JDBC driver for your DB to the app server classpath. We copied JDBC driver mysql-connector-java-5.1.5-bin.jar to [jboss-home]/server/default/lib

4. Configure CloverETL Server according to description in the next section (p. 16).

http://www.jboss.org/jbossas/downloads


16

5. Deploy WAR file

Copy clover.war to [jboss-home]/server/default/deploy

6. Start jboss via [jboss-home]/bin/run.sh (or run.bat on Windows OS)

It may take a couple of minutes until all applications are started.

7. Check JBoss response and CloverETL Server response

• JBoss administration console is accessible at http://localhost:8080/ by default. Default username/passwordis admin/admin

• CloverETL Server is accessible at http://localhost:8080/clover by default.

8. If you like, you can move default and example sandboxes (created automatically in the temp directory) to amore suitable directory on your filesystem.

• These sandboxes are created automatically during the first deployment and are located in the web-appdirectory, which is related to the specific deployment. If you redeployed the web application for a reason,the directory would be recreated. That is why it is better to move the sandboxes to a location which willnot change.

Configuration of CloverETL Server on JBoss

Default installation (without any configuration) is recommended only for evaluation purposes. For production,configurin at least tha DB data source and SMTP server is recommended.



• Create cloverServer.properties in a suitable directory

datasource.type=JNDIdatasource.jndiName=java:/CloverETLServerDSjdbc.dialect=org.hibernate.dialect.MySQLDialectlicense.file=/home/clover/config/license.dat

Do not change datasource.type and datasource.jndiName properties, but set a correct JDBCdialect according to your DB server (Chapter 18, Configuration (p. 96)). Also set path to your license file.

• Set system property (or environment property) clover_config_file.

It should contain the full path to the cloverServer.properties file created in the previous step.

The simplest way is seting java parameter in [jboss-home]/bin/run.sh, e.g.:

export JAVA_OPTS="$JAVA_OPTS -Dclover_config_file=/home/clover/config/cloverServer.properties"

Please do not override other settings in the JAVA_OPTS property. i.e. memory settings described above.

On Windows OS, edit [jboss-home]/bin/run.conf.bat and add this line to the section where optionsare passed to the JVM:


17

set JAVA_OPTS=%JAVA_OPTS% -Dclover_config_file=C:\JBoss6\cloverServer.properties

• This change requires restarting JBoss.




If you only have clover_license.war, extract it as a common zip archive and you will findlicense.dat in the WEB-INF subdirectory

2. Fill CloverETL Server parameter license.file with path to license.dat.

The best way how to configure license is setting the license.file property in thecloverServer.properties file as described in the previous section.

There are other ways how to do this. (See Chapter 18, Configuration (p. 96) for description of allpossibilities).

3. Changes in configuration require restarting the app-server.

Note

CloverETL license can be changed any time by replacing file license.dat. Then you have tolet CloverETL Server know the license is changed.


• Then click Reload license.

• Alternatively, you can restart CloverETL Server application.

Oracle WebLogic Server


1. Get CloverETL Server web archive file (clover.war) which is built for WebLogic.



• WebLogic (CloverETL Server is tested with 10.3.6, see http://www.oracle.com/technetwork/middleware/ias/downloads/wls-main-097127.html)

WebLogic has to be running and a domain has to be configured. You can check it by connecting toAdministration Console: http://hostname:7001/console/ (7001 is the default port for HTTP). Username andpassword are specified during installation.


It involves JVM parameters: -Xms -Xmx and -XX:MaxPermSize

See section the section called “Memory Settings” (p. 23) for details.


18

You can set it i.e. by adding

export JAVA_OPTIONS='$JAVA_OPTIONS -Xms128m -Xmx2048m -XX:MaxPermSize=512m' to the start script

This change requires restarting the domain.

3. Change HTTP Basic Authentication configuration

• When WebLogic finds "Authentication" header in an HTTP request, it tries to find a user in its own realm.His behavior has to be disabled so CloverETL could auhenticate users itself.

• Modify config file [domainHome]/config/config.xml. Add element: <enforce-valid-basic-auth-credentials>false</enforce-valid-basic-auth-credentials> intoelement <security-configuration> (just before the end tag).

4. Deploy WAR file (or application directory)

• Deploy the clover.war using the WebLogic Server Administration Console. See the Oracle FusionMiddleware Administrator's Guide (http://docs.oracle.com/cd/E23943_01/core.1111/e10105/toc.htm) fordetails.

5. Configure license (and other configuration properties)

• See separate section (p. 18) below

6. Check CloverETL Server URL

• Web-app is started automatically after deploy, so you can check whether it is up and running.

CloverETL Server is accessible at http://host:7001/clover by default. Port 7001 is the default WebLogicHTTP Connector port.

Configuration of CloverETL Server on Weblogic

Default installation (without any configuration) is recommended only for evaluation purposes. For production, atleast the DB data source and SMTP server configuration is recommended.



Create cloverServer.properties in a suitable directory.

Config file should contain DB datasource config, SMTP connection config, etc.

Set system property (or environment variable) clover_config_file pointing to the config properties file

• Set JAVA_OPTIONS variable in the WebLogic domain start script [domainHome]/startWebLogic.sh

JAVA_OPTIONS="${JAVA_OPTIONS} -Dclover_config_file=/path/to/clover-config.properties

• This change requires restarting Weblogic.



http://docs.oracle.com/cd/E23943_01/core.1111/e10105/toc.htm


19


If you only have clover_license.war, extract it as a common zip archive and you will findlicense.dat file in WEB-INF subdirectory

2. Fill CloverETL Server parameter license.file with path to license.dat file

The best way how to configure license, is setting property license.file in thecloverServer.properties file as described in the previous section.

There are other ways how to do this. (See Chapter 18, Configuration (p. 96) for description of allpossibilities).

3. Changes in configuration require restarting the app-server.

Note

Properly configured CloverETL license can be changed any time by replacing file license.dat.Then you have to let CloverETL Server know the license has changed.



• Or you can restart CloverETL Server application.

Possible issues during installation

Since CloverETL Server is considered a universal JEE application running on various application servers,databases and jvm implementations, problems may occur during the installation. These can be solved by a properconfiguration of the server environment. This section contains tips for the configuration.

Memory issues on Derby

If your server suddenly starts consuming too much resources (CPU, memory) despite having been working wellbefore, it might be beacuse of running the internal Derby DB. Typically, causes are incorrect/incomplete shutdownof Apache Tomcat and parallel (re)start of Apache Tomcat.

Solution: move to a standard (standalone) database.

How to fix this? Redeploy CloverETL Server:

1. Stop Apache Tomcat and verify there are no other instances running. If so, kill them.

2. Backup config.properties from webapps/clover/WEB-INF and clover/WEB-INF/sandboxes (if you have any data there).

3. Delete the webapps/clover directory.

4. Start Apache Tomcat server. It will automatically redeploy Clover Server.

5. Verify you can connect from Designer and from web.

6. Shutdown Apache Tomcat.

7. Restore config.properties and point it to your regular database.

8. Start Apache Tomcat.


20

JAVA_HOME or JRE_HOME environment variables are not defined

If you are getting this error message during an attempt to start your application server (mostly Tomcat), performthe following actions.

Linux:

These two commands will help you set paths to the variables on the server.

• [root@server /] export JAVA_HOME=/usr/local/java

• [root@server /] export JRE_HOME=/usr/local/jdk

As a final step, restart the application server.

Windows OS:

Set JAVA_HOME to your JDK installation directory, e.g. C:\Program Files\java\jdk1.6.0. Optionally,set also JRE_HOME to the JRE base directory, e.g. C:\Program Files\java\jre6.

Important

If you only have JRE installed, specify only JRE_HOME.

Tomcat log file catalina.out is missing on Windows

Tomcat start batch files for Windows aren't configured to create catalina.out file which contains standard output ofthe application. Catalinal.out may be vital when the Tomcat isn't started in console and any issue occurs. Or evenwhen Tomcat is executed in the console, it may be closed automatically just after the error message appears in it.

Please follow these steps to enable catalina.out creation:

• Modify [tomcat_home]/bin/catalina.bat. Add parameter "/B" to lines where "_EXECJAVA" variable is set.There should be two these lines. So they will look like this:

set _EXECJAVA=start /B [the rest of the line]Parameter /B causes, that "start" command doesn't open new console window, but runs the command it's ownconsole window.

• Create new startup file. e.g. [tomcat_home]/bin/startupLog.bat, containing only one line:

catalina.bat start > ..\logs\catalina.out 2<&1It executes Tomcat in the usual way, but standard output isn't put to the console, but to the catalina.out file.

Then use new startup file instead of [tomcat_home]/bin/startup.bat

Timeouts waiting for JVM

If you get the Jetty application server successfully running but cannot start Clover Server, it might be because ofthe wrapper waiting for JVM too long (it is considered a low-memory issue). Examine [JETTY_HOME]\logs\jetty-service.log for a line like this:

Startup failed: Timed out waiting for signal from JVM.

If it is there, edit [JETTY_HOME]\bin\jetty-service.conf and add these lines:


21

wrapper.startup.timeout=60wrapper.shutdown.timeout=60

If that does not help either, try setting 120 for both values. Default timeouts are 30 both.

clover.war as default context on Websphere (Windows OS)

If you are deploying clover.war on the IBM Websphere server without context path specified, be sure to checkwhether it is the only application running in the context root. If you cannot start Clover Server on Websphere,check the log and look for a message like this:

com.ibm.ws.webcontainer.exception.WebAppNotLoadedException: Failed to load webapp: Failed to load webapp: Context root /* is already bound. Cannot start application CloverETL

If you can see it, then this is the case. Getting rid of the issue, the easiest way is to stop all other (sample)applications and leave only clover.war running on the server. That should guarantee the server will be availablein the context root from now on (e.g. http://localhost:9080/).

Figure 2.2. Clover Server as the only running application on IBM Websphere

Tomcat 6.0 on Linux - Default DB

When using the internal (default) database on Linux, your Clover Server might fail on first start for no obviousreasons. Chances are that the /var/lib/tomcat6/databases directory was not created (because of accessrights in parent folders).

Solution: Create the directory yourself and try restarting the server. This simple fix was successfully tested withClover Server deployed as a WAR file via Tomcat web admin.

Derby.system.home cannot be accessed

If the server cannot start and the following message is in the log:



22

java.sql.SQLException: Failed to start database 'databases/cloverserver'

then see the next exception for details. After that check settings of the derby.system.home system property.It may point to an unaccessible directory, or files may be locked by another process. We suggest you set a specificdirectory as the system property.

Environment variables and more than one CloverETL Server instancesrunning on single machine

If you are setting environment variables like clover_license_file or clover_config_file ,remember you should not be running more than one CloverETL Server. Therefore if you ever needed to run moreinstances at once, use other ways of setting parameters (see Chapter 18, Configuration (p. 96) for descriptionof all possibilities) The reason is the environment variable is shared by all applications in use causing them to shareconfigurations and fail unexpectedly. Instead of environment variables you can use system properties (passed tothe application container process using parameter with -D prefix: -Dclover_config_file).

Special characters and slahes in path

When working with servers, you ought to stick to folder naming rules more than ever. Do not use any specialcharacters in the server path, e.g. spaces, accents, diacritics are all not recommended. It's unfortunatelly commonnaming strategy on Windows systems. It can produce issues which are hard to find. If you are experiencing weirderrors and cannot trace the source of them, why not install your application server in a safe destination like:

C:\JBoss6\

Similarly, use slashes but never backslahes in paths inside the *.properties files, e.g. when pointing to theClover Server license file. If you incorrectly use backlash, it will be considered an escape character and the servermay not work fine. This is an example of a correct path:

license.file=C:/CoverETL/Server/license.dat

JAXB and early versions of JVM 1.6

CloverETL Server contains jaxb 2.1 libraries since version 1.3. This may cause conflicts on early versions of JVM1.6 which contain jaxb 2.0. However JDK6 Update 4 release finally contains jaxb 2.1, thus update to this or newerversion of JVM solves possible conflicts.

File system permissions

Application server must be executed by OS user which has proper read/write permissions on file system. Problemmay occur, if app-server is executed by root user for the first time, so log and other temp files are created by rootuser. When the same app-server is executed by another user, it will fail because it cannot write to root's files.

JMS API and JMS third-party libraries

Missing JMS libraries do not cause fail of server startup, but it is issue of deployment on application server, thusit still suits to this chapter.

clover.war itself does not contain jms.jar, thus it has to be on application server's classpath. Most of the applicationservers have jms.jar by default, but i.e. tomcat does not. so if the JMS features are needed, the jms.jar has to beadded explicitly.


23

If "JMS Task" feature is used, there must be third-party libraries on server's classpath as well. The same approachis recommended for JMS Reader/Writer components, even if these components allow to specify external libraries.It is due to common memory leak in these libraries which causes "OutOfMemoryError: PermGen space".

Memory Settings

Current implementation of Java Virtual Machine allows only global configuration of memory for the JVM systemprocess. Thus whole application server, together with WARs and EARs running on it, share one memory space.

Default JVM memory settings is to low for running application container with CloverETL Server. Someapplication servers, like IBM Websphere, increase JVM defaults themselves, however they still may be too low.

The best memory limits depend on many conditions, i.e. transformations which CloverETL should execute. Pleasenote, that maximum limit isn't amount of permanently allocated memory, but limit which can't be exceeded. If thelimit was exhaused, the OutOfMemoryError would be raised.

You can set the minimum and maximum memory heap size by adjusting the "Xms" and "Xmx" JVM parameters.There are more ways how to change the settings depending on the used application container.

If you have no idea about the memory required for the transformations, a maximum of 1-2 GB heap memory isrecommended. This limit may be increased during transformations development when OutOfMemoryErroroccurs.

Memory space for loading classes (so called "PermGen space") is separated from heap memory, and can be setby the JVM parameter "-XX:MaxPermSize". By default, it is just 64 MB which is not enough for enterpriseapplications. Again, suitable memory limit depends on various criteria, but 512 MB should be enought in mostcases. If the PermGen space maximum is too low, OutOfMemoryError: PermGen space may occur.

Please see the specific container section for details how to make the settings.

Upgrading Server to Newer Version

Getting New Version to Work

1. Get the web archive file (WAR) with a newer build of CloverETL Server.

2. Re-deploy the web application. Instructions how to do that are application server dependant - see EnterpriseServer (p. 4) for installation details on all supported servers. After you re-deploy, your new server will beconfigured based on the previous version's configuration.

3. If any changes to the database schema are necessary, the new server will automatically make them when yourun it for the first time. It's recommended to backup database before upgrade.

Upgrading Server License

1. The license file is shipped as a text containing a unique set of characters. If you:

• received the new license as a file (*.dat), then simply overwrite your old license file.

• have been sent the licence text e.g inside an e-mail, then copy the license contents (i.e. all text betweenCompany and END LICENSE) into a new file called clover-license.dat. Next, overwrite the oldlicense file with the new one.

2. In Clover Server configuration, change the full path to your new license file if necessary.

3. In server web GUI →Monitoring →License, click Reload license. Alternatively, restart Clover Server.


24

Important

Evaluation Version - a mere upgrade of your license is not sufficient. When moving from evaluationto enterprise server, you should not use the default configuration and database. Instead, take sometime to configure Clover Server so that it best fits your production environment.

25

Chapter 3. Server Side Job files - SandboxesSandbox is a base storage unit for project. Sandbox is actually a server-side analogy to a CloverETL Designerproject. Since CloverETL Designer has a connector to CloverETL Server, a designer project and a server sandboxmay be linked together. This remote CloverETL Designer project looks and works like common local project, butall files are stored on the server side and all operations are performed on server side. See CloverETL Designermanual for details on configuring a connection to the server.

Technically speaking, a sandbox is a dedicated directory on the server file system. A sandbox cannot containanother sandbox. It is recommended to have one directory as sandboxes container and create a subdirectory foreach sandbox. Files and directories in sandboxes are read by JVM of Application Server. Thus, all these directoriesmust be accessible to the OS user who executes JVM of Application Server. i.e. If Apache Tomcat is executed asan OS service by "tomcat" user, all sandboxes must be accessible to this user.

In cluster mode, there are three sandbox types: "shared", "local" and "partitioned". See Chapter 24,Clustering (p. 115) for details.

Figure 3.1. Sandboxes Section in CloverETL Server Web GUI

Each sandbox is defined by following attributes:

Table 3.1. Sandbox attributes

ID Unique "name" of the sandbox. It is used in server APIs to identify sandbox. It must meetcommon rules for identifiers. It is specified by user in during sandbox creation and it can bemodified later. Note: modifying is not recommended, because it may be already used by someCS APIs clients.

Name Sandbox name used just for display. It is specified by user in during sandbox creation and itcan be modified later.

Root path Absolute server side file system path to sandbox root. It is specified by user during sandboxcreation and it can be modified later. This attribute is used only in standalone mode. See Chapter24, Clustering (p. 115) for details about cluster mode.

Owner It is set automatically during sandbox creation. It may be modified later.

Chapter 3. Server SideJob files - Sandboxes

26

Figure 3.2. Sandbox Detail in CloverETL Server Web GUI

Referencing files from the ETL graph or Jobflow

In some components you can specify file URL attribute as a reference to some resource on the file system. Alsoexternal metadata, lookup or DB connection definition is specified as reference to some file on the filesystem.With CloverETL Server there are more ways how to specify this relation.

• Relative path

All relative paths in your graphs are considered as relative paths to the root of the same sandbox which containsjob file (ETL graph or Jobflow).

• sandbox:// URLs

Sandbox URL allows user to reference the resource from different sandboxes with standalone CloverETL Serveror the cluster. In cluster environment, CloverETL Server transparently manages remote streaming if the resourceis accessible only on some specific cluster node.

See Using a Sandbox Resource as a Component Data Source (p. 119) for details about the sandbox URLs.

Sandbox Content Security and Permissions

Each sandbox has its owner which is set during sandbox creation. This user has unlimited privileges to this sandboxas well as administrators. Another users may have access according to sandbox settings.


27

Figure 3.3. Sandbox Permissions in CloverETL Server Web GUI

Permissions to specific sandbox are modifiable in Permissions tab in sandbox detail. In this tab, selected usergroups may be allowed to perform particular operations.

There are 3 types of operations:

Table 3.2. Sandbox permissions

R - read Users can see this sandbox in their sandboxes list.

W - write Users can modify files in the sandbox through CS APIs.

X - execution Users can execute graphs in this sandbox. Note: graph executed by "graph event listener" isactually executed by the same user as graph which is source of event. See details in "graphevent listener". Graph executed by schedule trigger is actually executed by the scheduleowner. See details in Chapter 6, Scheduling (p. 45).

Please note that, these permissions modify access to the content of specific sandboxes. In additions, it's possible toconfigure permissions to perform operations with sandbox configuration. e.g. create sandbox, edit sandbox, deletesandbox, etc. Please see Chapter 5, Users and Groups (p. 38) for details.

Sandbox Content

Sandbox should contain jobflows, graphs, metadata, external connection and all related files. Files especially graphor jobflow files are identified by relative path from sandbox root. Thus you need two values to identify specificjob file: sandbox and path in sandbox. Path to the Jobflow or ETL graph is often referred as "Job file".


28

Figure 3.4. Web GUI - section "Sandboxes" - context menu on sandbox

Although web GUI section sandboxes isn't file-manager, it offers some useful features for sandbox management.

Figure 3.5. Web GUI - section "Sandboxes" - context menu on folder

Download sandbox in ZIP

Select sandbox in left panel, then web GUI displays button "Download sandbox in ZIP" in the tool bar on theright side.

Created ZIP contains all readable sandbox files in the same hierarchy as on file system. You can use this ZIP filefor upload files to the same sandbox, or another sandbox on different server instance.


29

Figure 3.6. Web GUI - download sandbox in ZIP

Upload ZIP to sandbox

Select sandbox in left panel. You must have write permission to the selected sandbox. Then select tab "UploadZIP" in the right panel. Upload of ZIP is parametrized by couple of switches, which are described below. Opencommon file chooser dialog by button "+ Upload ZIP". When you choose ZIP file, it is immediately uploaded tothe server and result message is displayed. Each row of the result message contains description of one single fileupload. Depending on selected options, file may be skipped, updated, created or deleted.


30

Figure 3.7. Web GUI - upload ZIP to sandbox

Figure 3.8. Web GUI - upload ZIP results

Table 3.3. ZIP upload parameters

Label Description

Encoding of packed filenames

File names which contain special characters (non ASCII) are encoded. By this selectbox, you choose right encoding, so filenames are decoded properly.

Overwrite existing files If this switch is checked, existing file is overwriten by new one, if both of them arestored in the same path in the sandbox and both of them have the same name.

Replace sandbox content If this option is enabled, all files which are missing in uploaded ZIP file, but theyexist in destination sandbox, will be deleted. This option might cause loose of data, souser must have special permission "May delete files, which are missing in uploadedZIP" to enable it.


31

Download file in ZIP

Select file in left panel, then web GUI displays button "Download file in ZIP" in the tool bar on the right side.

Created ZIP contains just selected file. This feature is useful for large files (i.e. input or output file) which cannotbe displayed directly in web GUI. So user can download it.

Figure 3.9. Web GUI - download file in ZIP

Download file HTTP API

It is possible to download/view sandbox file accessing "download servlet" by simple HTTP GET request:

http://[host]:[port]/[Clover Context]/downloadFile?[Parameters]

Server requires BASIC HTTP Authentication. Thus with linux command line HTTP client "wget" it would looklike this:

wget --user=clover --password=clover http://localhost:8080/clover/downloadFile?sandbox=default\&file=data-out/data.dat

Please note, that ampersand character is escaped by back-slash. Otherwise it would be interpreted as command-line system operator, which forks processes.

URL Parameters

• sandbox - Sandbox code. Mandatory parameter.

• file - Path to the file relative from sandbox root. Mandatory parameter.

• zip - If set to "true", file is returned as ZIP and response content type is "application/x-zip-compressed". Bydefault it is false, so response is content of the file.


32

Job config properties

Each ETL graph or Jobflow may have set of config properties, which are applied during the execution. Propertiesare editable in web GUI section "sandboxes". Select job file and go to tab "Config properties".

The same config properties are editable even for each sandbox. Values specified for sandbox are applied for eachjob in the sandbox, but with lower priority then config properties specified for the job.

If neither sandbox or job have config properties specified, defaults from main server configuration areapplied. Global config properties related to Job config properties have prefix "execution.". E.g. server property"executor.classpath" is default for Job config property "classpath". (See Chapter 18, Configuration (p. 96)for details)

In addition, it is possible to specify additional job parameters, which can be used as placeholders in job XML.Please keep in mind, that these placeholders are resolved during loading and parsing of XML file, thus such jobcouldn't be pooled.


33

Table 3.4. Job config parameters

Property name Default value Description

tracking_interval 2000 Interval in ms for sampling nodes status in runningtransformation.

max_running_concurrently unlimited Max number of concurrently running instances ofthis transformation.

enqueue_executions false Boolean value. If it is true, executions abovemax_running_concurrently are enqueued, if it isfalse executions above max_running_concurrentlyfail.

log_level INFO Log4j log level for this graph executions. (ALL| TRACE | DEBUG | INFO | WARN | ERROR| FATAL) For lower levels (ALL, TRACE orDEBUG), also root logger level must be setto lower level. Root logger log level is INFOby default, thus transformation run log does notcontain more detail messages then INFO event ifjob config parameter "log_level" is set properly.See Chapter 21, Logging (p. 110) for detailsabout log4j configuration.

max_graph_instance_age 0 Time interval in ms which specifies how longmay transformation instance last in server's cache.0 means that transformation is initialized andreleased for each execution. Transformation cannotbe stored in the pool and reused in somecases (transformation uses placeholders usingdynamically specified parameters)

classpath List of paths or jar files which containexternal classes used in the job file(transformations, generators, JMS processors).Separator is specified by Engine property"DEFAULT_PATH_SEPARATOR_REGEX".Directory path must always end with slashcharacter "/", otherwise ClassLoader doesn'trecognize it's a directory. Server alwaysautomatically adds "trans" subdirectory of job'ssandbox, so It doesn't have to be added explicitly.

skip_check_config default value is taken fromengine property

Switch which specifies whether check config mustbe performed before transformation execution.

password Password for decoding of encoded DB connectionpasswords.

verbose_mode true If true, more descriptive logs of job runs aregenerated.

use_jmx true If true, job executor registers jmx mBean of runningtransformation.

debug_mode false If true, edges with enabled debug store datainto files in debug directory. See property"graph.debug_path".

Without explicit setting, running of a graph fromDesigner with server integration would set thedebug_mode to true. On the other hand, runningof a graph from the server console sets thedebug_mode to false.

executor.use_local_context_urlIf true, the context URL ofa running job will be a local"file:" URL. Otherwise, a"sandbox:" URL will be used.

false

executor.jobflow_token_trackingIf false, token tracking injobflow executions will bedisabled.

true


34

Figure 3.10. Job config properties

35

Chapter 4. Viewing Job Runs - Executions HistorySection Executions History shows all persistent job executions. Table shows only basic info about the job: RunId, Job file, current status, time of execution, and some useful links.

Filtering and ordering

User may filter records in the table by various criteria: Run ID, Job File, Date/time of execution, Status, userwho executed the job. Also user may show children executions, which are filtered-out by default (e.g. workers ofpartitioned executions or jobs executed from jobflows).

Last jobs are on the top by default.

Please note, that some jobs may have disabled persistence to increase performance. It would be typically jobsexecuted by Launch Services, since the performance may be more important then detail info about the execution.

Figure 4.1. Executions History - executions table

When some job execution is selected in the table, the detail info is shown on the right side.

Chapter 4. Viewing JobRuns - Executions History

36

Table 4.1. Persistent run record attributes

Attribute Description

Run ID Unique number identifying the run of the job. Server APIs usually return this numberas simple response of execution request. It's useful as parameter of subsequent callsfor specification of the job execution.

Execution type Type of job as recognized by the server. STANDALONE for ETL graph, JOBFLOWfor Jobflow, MASTER for main record of partitioned execution in cluster,PARTITIONED_WORKER for worker record of partitioned execution in cluster

Parent run ID Run ID of the parent job. Typically Jobflow which executed this job, or masterexecution which encapsulate this worker execution.

Root run ID Run ID of the root parent job. Job execution which wasn't executed by another parentjob.

Nested jobs Indication that this job execution has or has not any child execution.

Node In cluster mode shows ID of the cluster node which this execution was running on.

Executed by User which executed the job. Either directly using some API/GUI or indirectly usingthe scheduling or event listeners.

Sandbox Sandbox containing job file. For jobs which are sent together with execution request,so the job file doesn't exist on the server site, it's set to "default" sandbox.

Job file Path to job file, relative to the sandbox root. For jobs which are sent together withexecution request, so the job file doesn't exist on the server site, it's set to generatedstring.

Job version Revision of the job file. It's string generated by CloverETL Designer and stored inthe job file.

Status Status of the job execution. READY - waiting for execution start, RUNNING -processing job, FINISHED OK - job finished without any error, ABORTED - jobwas aborted directly using some API/GUI or by parent Jobflow, ERROR - job failed,N/A (not awailable) - server process died suddenly, so it couldn't properly abort thejobs, so after sertart the jobs with unknown status are set as N/A

Started Server date-time (and timezone) of the execution start.

Finished Server date-time (and timezone) of the execution finish.

Duration Execution duration

Error in component ID If the job failed due the error in a component, this field contains ID of the component.

Error in component type If the job failed due the error in a component, this field contains type of thecomponent.

Error message If the job failed, this field contains error description.

Exception If the job failed, this field contains error stack trace.

Input parameters List of input parameters passed to the job. Job file can't be cached, since theparameters are applied during loading from the job file. Job file isn't cached bydefault.

Input dictionary List of dictionary elements passed to the job. Dictionary is used independently onjob file caching.

Output dictionary List of dictionary elements at the moment the job ends.

For jobs which have some children executions, e.g. partitioned or jobflows also executions hierarchy tree is shown.

Chapter 4. Viewing JobRuns - Executions History

37

Figure 4.2. Executions History - overall perspective

Since the detail panel and expecially job logs may be wide, it may be usefull to hide table on the left, so the detailpanel spreads. Click on the link "Hide results" on the top of the list panel to hide panel. Then to show list panelagan, click to the "Show results" button on the left.

Figure 4.3. Executions Hierarchy with docked list of jobs

Executions hierarchy may be rather complex, so it's possible to filter the content of the tree by fulltext filter.However when the filter is used, the selected executions aren't hierarchically structured.

38

Chapter 5. Users and GroupsCloverETL Server implements security module, which manages users and groups. Security module may beglobally switched off (see Chapter 18, Configuration (p. 96) for details), but by default it is on, and all interfacesrequire client authentication by username and password. Relation between users and groups is N:M, thus one usermay be assigned in more groups and one group may be assigned in more users.

All relations between users and groups are configurable in web GUI in sections Users and Groups.

Both sections are accessible only for users which have "List users" ("List groups" resp.) permission. To modifyusers/groups "create", "edit" and "delete" permissions are necessary.

LDAP authentication

Since 3.2 it's possible to configure CloverETL Server to use LDAP server for users authentication. So thecredentials of users registered in LDAP may be used for authentication to any CloverETL Server interface (APIor GUI).

However authorization (access levels to sandboxes content and privileges for operations) is still handled by Cloversecurity module. Each user, event though logged-in using LDAP authentication, must have his own "user" record(with related groups) in CloverETL security module. So there must be the user with the same username and domainset to "LDAP". If no such user record exists, it's automatically created according to CloverETL configuration.

What does the CloverETL do to authenticate a LDAP user?

1. User specifies the LDAP credentials i.e. in login form to the web GUI

2. CloverETL Server connects to the LDAP and checks whether the user exists (it uses specified search to lookupin LDAP)

3. If the user exists in LDAP, CloverETL Server performs authentication

4. If succeeded, CloverETL Server searches for LDAP user's groups.

5. CloverETL Server checks whether the user is assigned in LDAP groups which are allowed to login to Clover.

6. Clover user record is created/updated according to current LDAP values.

7. Clover user is assigned to the Clover groups according to his current assignation to the LDAP groups.

8. User is logged-in

Note

Switching domains:

• If a user was created as LDAP and then switched to clover domain, you have to set a passwordfor him in Change password tab.

• If a user was created as clover and then switched to LDAP domain, he has a password inclover domain, but it is overriden by the LDAP password. After switching back to clover domain,the original password is re-used. It can be reset in the Change password tab if needed (e.g.forgotten).

Configuration

By default CloverETL Server allows only its own internal mechanism for authentication. To enable authenticationwith LDAP, set config property "security.authentication.allowed_domains" properly. It's list of user domainswhich are used for authentication.

Chapter 5. Users and Groups

39

Currently there are 2 authentication mechanism implemented: "LDAP" and "clover" ("clover" is identifier ofCloverETL internal authentication and may be changed by security.default_domain property, but only for white-labelling purposes). To enable LDAP authentication, set value to "LDAP" (only LDAP) or "clover,LDAP". Usersfrom both domain may login. It's recommended to allow both mechanisms together, until the LDAP is properlyconfigured. So the admin user can still login to web GUI although the LDAP connection isn't properly configured.

Basic LDAP connection properties

# Implementation of context factorysecurity.ldap.ctx_factory=com.sun.jndi.ldap.LdapCtxFactory# timeout for all queries sent to LDAP serversecurity.ldap.timeout=5000# limit for number of records returned from LDAPsecurity.ldap.records_limit=50

# URL of LDAP serversecurity.ldap.url=ldap://hostname:port# Some generic UserDN which allows lookup for the user and groups. security.ldap.userDN=# Password for the user specified abovesecurity.ldap.password=

Configuration of user lookup

Specified values work for this specific LDAP tree:

• dc=company,dc=com

• ou=groups

• cn=admins(objectClass=groupOfNames,member=(uid=smith,dc=company,dc=com),member=(uid=jones,dc=company,dc=com))

• cn=developers (objectClass=groupOfNames,member=(uid=smith,dc=company,dc=com))

• cn=consultants (objectClass=groupOfNames,member=(uid=jones,dc=company,dc=com))

• ou=people

• uid=smith (fn=John,sn=Smith,[email protected])

• uid=jones (fn=Bob,sn=Jones,[email protected])

Following properties are necessary for lookup for the LDAP user by his username. (step [2] in the login processabove)

# Base specifies the node of LDAP tree where the search startssecurity.ldap.user_search.base=dc=company,dc=eu# Filter expression for searching the user by his username. # Please note, that this search query must return just one record.# Placeholder ${username} will be replaced by username specified by the logging user.security.ldap.user_search.filter=(uid=${username})# Scope specifies type of search in "base". There are three possible values: SUBTREE | ONELEVEL | OBJECT # http://download.oracle.com/javase/6/docs/api/javax/naming/directory/SearchControls.htmlsecurity.ldap.user_search.scope=SUBTREE

Following properties are names of attributes from the search defined above. They are used for getting basic infoabout the LDAP user in case the user record has to be created/updated by Clover security module: (step [6] inthe login process above)


40

security.ldap.user_search.attribute.firstname=fnsecurity.ldap.user_search.attribute.lastname=snsecurity.ldap.user_search.attribute.email=mail# This property is related to the following step "searching for groups". # Groups may be obtained from specified user's attribute, or found by filter (see next paragraph)# Please leave this property empty if the user doesn't have such attribute. security.ldap.user_search.attribute.groups=memberOf

In the following step, clover tries to find groups which the user is assigned to. (step [4] in the login process above).There are two ways how to get list of groups which the user is assigned to. The user-groups relation is specifiedon the "user" side. The user record has some attribute with list of groups. It's "memberOf" attribute usually. Or therelation is specified on the "group" side. The group record has attribute with list of assigned users. It's "member"attribute usually.

In case the relation is specifien on users side, please specify property:

security.ldap.user_search.attribute.groups=memberOf

Leave it empty otherwise.

In case the relation is specifien on groups side, please specify properties for searching:

security.ldap.groups_search.base=dc=company,dc=com# Placeholder ${userDN} will be replaced by user DN found by the search above# If the filter is empty, searching will be skipped.security.ldap.groups_search.filter=(&(objectClass=groupOfNames)(member=${userDN}))security.ldap.groups_search.scope=SUBTREE

Otherwise, please leave property security.ldap.groups_search.filter empty, so the search will be skipped.

Clover user record will be assigned to the clover groups according to the LDAP groups found by the search (orthe attribute). (Groups check is performed during each login)

# Value of the following attribute will be used for lookup for the Clover group by its code.# So the user will be assigned to the Clover group with the same "code"security.ldap.groups_search.attribute.group_code=cn

It's also possible to specify LDAP groups which are able to login to Clover. (step [5] in the login process above)

# Semicolon separated list of LDAP group DNs (distinguished names). # LDAP user must be assigned to one or more of these groups, otherwise new clover user can't be created.# Special value "_ANY_" disables this check and basically any LDAP user may login.# If the LDAP group DNs are configured, also security.ldap.groups_search.* properties must be configured.# value could be e.g. "cn=developers,dc=company,dc=com;cn=admins,dc=company,dc=com"security.ldap.allowed_ldap_groups=_ANY_

Web GUI section Users

This section is intended to users management. It offers features in dependence of user's permissions. i.e. User mayenter this section, but cannot modify anything. Or user may modify, but cannot create new users.

All possible features of users section:


41

• create new user

• modify basic data

• change password

• disable/enable user

• assign user to groups - Assignment to groups gives user proper permissions

Table 5.1. After default installation above empty DB, there are two users created

User name Description

clover Clover user has admin permissions, thus default password "clover"should be changed after installation.

system System user is used by application instead of common user, whenno other user can be used. i.e. when security is globally switchedoff. This user cannot be removed and it is impossible to login asthis user.

Figure 5.1. Web GUI - section "Users" under "Configuration"

Table 5.2. User attributes


username Common user identifier. Must be unique, cannot contain spaces orspecial characters, just letters and numbers.

password Case sensitive password. If user looses his password, the new onemust be set. Password is stored in encrypted form for securityreasons, so it cannot be retrieved from database and must bechanged by the user who has proper permission for such operation.

first name

last name

email Email which may be used by CloverETL administrator or byCloverETL server for automatic notifications. See Task - SendEmail (p. 57) for details.

Edit user record

User with permission "Create user" or "Edit user" can use this form to set basic user parameters.


42

Figure 5.2. Web GUI - edit user

Change users Password

If user looses his password, the new one must be set. So user with permission "Change passwords" can use thisform to do it.

Figure 5.3. Web GUI - change password

Group assignment

Assignment to groups gives user proper permissions. Only logged user with permission "Groups assignment" canaccess this form and specify groups which the user is assigned in. See Web GUI section Groups (p. 43) fordetails about permissions.

Figure 5.4. Web GUI - groups assignment


43

Disabling / enabling users

Since user record has various relations to the logs and history records, it can't be deleted. So it's disabled instead.It basically means, that the record doesn't display in the list and the user can't login.

However disabled user may be enabled again. Please note, that disabled user is removed from its groups, so groupsshould be assigned properly after re-enabling.

Web GUI section Groups

Group is abstract set of users, which gives assigned users some permissions. So it is not necessary to specifypermission for each single user.

There are independent levels of permissions implemented in CloverETL Server

• permissions to Read/Write/eXecute in sandboxes - sandbox owner can specify different permissions for differentgroups. See Sandbox Content Security and Permissions (p. 26) for details.

• permissions to perform some operation - user with operation permission "Permission assignment" may assignspecific permission to existing groups.

• permissions to launch specific service - see Chapter 17, Launch Service (p. 89) for details.

Table 5.3. Default groups created during installation

Group name Description

admins This group has operation permission "all" assigned, which means,that it has unlimited permission. Default user "clover" is assignedto this group, which makes him administrator.

all users Every single CloverETL user is assigned to this group by default.It is possible to remove user from this group, but it is not arecommended approach. This group is useful for some permissionsto sandbox or some operation, which you would like to makeaccessible for all users without exceptions.

Figure 5.5. Web GUI - section "Groups"

Users Assignment

Relation between users and groups is N:M. Thus in the same way, how groups are assignable to users, users areassignable to groups.


44

Figure 5.6. Web GUI - groups assignment

Groups permissions

Groups permissions are structured as tree, where permissions are inherited from root to leafs. Thus if somepermission (tree node) is enabled (blue dot), all permissions in sub tree are automatically enabled (white dot).Permissions with red cross are disabled.

Thus for "admin" group just "all" permission is assigned, every single permission in sub tree is assignedautomatically.

Figure 5.7. Tree of permissions

45

Chapter 6. SchedulingScheduling allows user to create his own timetable for operations which he does not want to trigger manually. Eachschedule represents separated timetable and basically its specification WHEN to do something and WHAT to do.

In cluster environment, scheduling is processed only on master node, thus tasks are triggered only on master node.

Figure 6.1. Web GUI - section "Scheduling" - create new

Timetable Setting

This section should describe how to specify WHEN schedule should be triggered. Please keep in mind, that exacttrigger times are not guaranteed. There may be couple of seconds delay. Schedule itself can be specified in differentways.

• Onetime Schedule (p. 45)

• Periodical schedule by Interval (p. 46)

• Periodical schedule by timetable (Cron Expression) (p. 47)

Onetime Schedule

It is obvious, that this schedule is triggered just once.

Table 6.1. Onetime schedule attributes

Type "onetime"

Start date/time Date and time, specified with minutes precision.

Chapter 6. Scheduling

46

Figure 6.2. Web GUI - onetime schedule form

Figure 6.3. Web GUI - schedule form - calendar

Periodical schedule by Interval

This type of schedule is the simplest periodical type. Trigger times are specified by these attributes:


47

Table 6.2. Periodical schedule attributes

Type "periodic"

Periodicity "interval"


End date/time Date and time, specified with minutes precision.

Interval in minutes Specifies interval between two trigger times. Next task is triggered even ifprevious task is still running.

Fire misfired ASAP switch If checked and trigger time is missed because of any reason (i.e. serverrestart), it will be triggered immediately, when it is possible. Otherwise itis ignored and it will be triggered at next scheduled time.

Figure 6.4. Web GUI - periodical schedule form

Periodical schedule by timetable (Cron Expression)

Timetable is specified by powerful (but a little bit tricky) cron expression.

Table 6.3. Cron periodical schedule attributes

Type "periodic"

Periodicity "interval"


End date/time Date and time, specified with minutes precision.

Cron expression Cron is powerful tool, which uses its own format for scheduling. Thisformat is well known among UNIX administrators. i.e. "0 0/2 4-23 * * ?"means "every 2 minutes between 4:00am and 11:59pm".

Fire misfired ASAP switch If checked and trigger time is missed because of any reason (i.e. serverrestart), it will be triggered immediately when it is possible. Otherwise it isignored and it will be triggered at next scheduled time.


48

Figure 6.5. Cron periodical schedule form

Tasks

Task basically specifies WHAT to do at trigger time. There are several tasks implemented for schedule and forgraph event listener as follows:

• Task - Execution of Graph (p. 48)

• Task - Execution of Jobflow (p. 49)

• Task - Kill Job (p. 50)

• Task - Execution of Shell Command (p. 51)

• Task - Send Email (p. 52)

• Task - Execute Groovy Code (p. 52)

• Task - Archive Records (p. 53)

Task - Execution of Graph

Please note that behaviour of this task type is almost the same as Task - Execution of Jobflow (p. 49)


49

Table 6.4. Attributes of "Graph execution" task

Task type "execute graph"

Sandbox This select box contains sandboxes which are readable for logger user.Select sandbox which contains graph to execute.

Graph This select box is filled by all graphs files accessible in selected sandbox.

Parameters Key-value pairs which are passed to the executed job as parameters.Besides, if this task is triggered by job (graph or jobflow) event, you canspecify source job parameters, which shall be passed from the source jobto executed job. i.e. event source has these parameters: paramName2 withvalue "val2", paramName3 with value "val3", paramName5 with value"val5". Task has "Parameters" attribute set like this:

paramName1=paramValue1 paramName2= paramName3 paramName4

So executed job gets these parameters and values: paramName1 withvalue "paramValue1" (specified explicitly in the task configuration)paramName2 with value "" (empty string specified explicitly in the taskconfiguration overrides event source parameters), paramName3 with value"val3" (value is taken from event source). These parameters aren't passed:paramName4 isn't passed, since it does not have any value in event source.paramName5 isn't passed, since it is not specified among the parameters topass in the task.

Event parameters like "EVENT_RUN_RESULT", "EVENT_RUN_ID"etc. are passed to the executed job without limitations.

Figure 6.6. Web GUI - Graph execution task

Task - Execution of Jobflow

Please note that behaviour of this task type is almost the same as Task - Execution of Graph (p. 48)


50

Table 6.5. Attributes of "Jobflow execution" task

Task type "execute jobflow"

Sandbox This select box contains sandboxes which are readable for logger user.Select sandbox which contains jobflow to execute.

Jobflow This select box is filled by all jobflow files accessible in selected sandbox.

Parameters Key-value pairs which are passed to the executed job as parameters.Besides, if this task is triggered by job (graph or jobflow) event, you canspecify source job parameters, which shall be passed from the source jobto executed job. i.e. event source has these parameters: paramName2 withvalue "val2", paramName3 with value "val3", paramName5 with value"val5". Task has "Parameters" attribute set like this:

paramName1=paramValue1 paramName2= paramName3 paramName4

So executed job gets these parameters and values: paramName1 withvalue "paramValue1" (specified explicitly in the task configuration)paramName2 with value "" (empty string specified explicitly in the taskconfiguration overrides event source parameters), paramName3 with value"val3" (value is taken from event source). These parameters aren't passed:paramName4 isn't passed, since it does not have any value in event source.paramName5 isn't passed, since it is not specified among the parameters topass in the task.

Event parameters like "EVENT_RUN_RESULT", "EVENT_RUN_ID"etc. are passed to the executed job without limitations.

Figure 6.7. Web GUI - Jobflow execution task

Task - Kill Job

This task, when activated kills/aborts specified job (ETL graph or jobflow), if it is currently running.


51

Table 6.6. Attributes of "Kill Job" task

Task type "kill job"

Kill source of event If this switch is on, task will kill job which is source of the event, whichactivated this task. Attributes sandbox and job are ignored.

Sandbox Select sandbox which contains job to kill. This attribute works only when"Kill source of event" switch is off.

Job This select box is filled by all jobs accessible in selected sandbox. Allinstances of selected job, whose are currently running will be killed. Thisattribute works only when "Kill source of event" switch is off.

Figure 6.8. Web GUI - "Kill job"

Task - Execution of Shell Command

Table 6.7. Attributes of "Shell command" task

Task type "shell command"

Command line Command line for execution of external process.

Working directory Working directory for process. If not set, working directory of applicationserver process is used.

Timeout Timeout in milliseconds. After period of time specified by this number,external process is terminated and all results are logged.

Figure 6.9. Web GUI - shell command


52

Task - Send Email

This task is very useful, but for now only as response for graph events. This feature is very powerful for monitoring.(see Chapter 7, Graph Event Listeners (p. 55) for description of this task type).

Note: It seems useless to send emails periodically, but it may send current server status or daily summary. Thesefeatures will be implemented in further versions.

Task - Execute Groovy Code

This type of task allows execute code written in script language Groovy. It is possible to use some variables. Onlyparameter of this task is source code of written in Groovy.

Table 6.8. List of variables available in Groovy code

variable class description availability

event com.cloveretl.server.events.AbstractServerEvent every time

task com.cloveretl.server.persistent.Task every time

now java.util.Date current time every time

parametersjava.util.Properties Properties of task every time

user com.cloveretl.server.persistent.UserSame as event.getUser() every time

run com.cloveretl.server.persistent.RunRecord When the event is instance ofGraphServerEvent

tracking com.cloveretl.server.persistent.TrackingGraphsame as run.getTrackingGraph() When the event is instance ofGraphServerEvent

sandbox com.cloveretl.server.persistent.Sandboxsame as run.getSandbox() When the event is instance ofGraphServerEvent

schedulecom.cloveretl.server.persistent.Schedulesame as((ScheduleServerEvent)event).getSchedule()

When the event is instance ofScheduleServerEvent

servletContextjavax.servlet.ServletContext every time

cloverConfigurationcom.cloveretl.server.spring.CloverConfigurationConfiguration values forCloverETL Server

every time

serverFacadecom.cloveretl.server.facade.api.ServerFacadeReference to the facadeinterface. Useful for callingCloverETL Server core.

WAR file contains JavaDoc offacade API and it is accessibleon URL: http://host:port/clover/javadoc/index.html

every time

sessionTokenString Valid session token of the userwho owns the event. It is usefulfor authorisation to the facadeinterface.

every time

Variables run, tracking and sandbox are available only if event is instance of GraphServerEvent class. Variableschedule is available only for ScheduleServerEvent as event variable class.

Example of use Groovy script

This example shows script which writes text file describing finished graph. It shows use of 'run' variable.

../javadoc/index.html



53

import com.cloveretl.server.persistent.RunRecord;String dir = "/tmp/";RunRecord rr = (RunRecord)run;

String fileName = "report"+rr.getId()+"_finished.txt";

FileWriter fw = new FileWriter(new File(dir+fileName));fw.write("Run ID :"+rr.getId()+"\n");fw.write("Graph ID :"+rr.getGraphId()+"\n");fw.write("Sandbox :"+rr.getSandbox().getName()+"\n");fw.write("\n");fw.write("Start time :"+rr.getStartTime()+"\n");fw.write("Stop time :"+rr.getStopTime()+"\n");fw.write("Duration :"+rr.getDurationString()+"\n");fw.write("Final status :"+rr.getFinalStatus()+"\n");fw.close();

Task - Archive Records

As name suggests, this task can archive (or delete) obsolete records from DB.

Table 6.9. Attributes of "archive records" task

Task type "archivator"

Older then Time period (in minutes) - it specifies which records are evaluated asobsolete. Records older then the specified interval are stored in archives.

Archivator type There are two possible values: "archive" or "delete". Delete removesrecords without any possibility of UNDO operation. Archive removesrecords from DB, but creates ZIP package with CSV files containingdeleted data.

Output path for archives This attribute makes sense only for "archive" type.

Include executions history

Run record with status If status is selected, only run records with specified status will be archived.It is useful e.g. If you want to delete records for successfully finished jobs,but you want to keep failed jobs for further investigation.

Include temp files

Temp files with record status If status is selected, only temp files related to the run records with selectedstatus will be archived. It is useful e.g. If you want to delete files forsuccessfully finished jobs, but you want to keep failed jobs for furtherinvestigation.

Include tasks history If checked, archivator will include run records. Log files of graph runs areincluded as well.

Task types If this task type is selected, only logs for selected task type are archived.

Task result mask Mask applied to task log result attribute. Only records whose result meetsthis mask are archived. Specify string without any wildcards. Each tasklog which contains specified string in the "result" attribute will be deleted/archived. Case sensitivity depends on database collation.

Include debug files If checked, archivator removes all graph debug files older then giventimestamp defined in "Older than" attribute.

Include dictionary files If checked, archivator removes all dictionary temporary files older thengiven timestamp defined in "Older than" attribute.


54

Figure 6.10. Web GUI - archive records

55

Chapter 7. Graph Event ListenersGraph event listener is powerful feature, which allows user to monitor success of failure of ETL graph executions.It is also possible to create relations between executions, or execute backup script in dependence of graph successor failure.

Please note that Graph Event Listeners work very similar to Jobflow Event Listeners (Chapter 8, Jobflow EventListeners (p. 63)) in many ways, since ETL Graph and Jobflow are both "jobs" from the point of view ofCloverETL Server.

In cluster environment, event exists only on cluster node, which runs jobflow, thus if the node isn't explicitlyspecified, the task is triggered on the same node.

Graph Events

Each event carries properties of graph, which is source of event. If there is a event listener specified, task may usethese properties. i.e. next graphs in the chain may use "EVENT_FILE_NAME" placeholder which activated firstgraph in the chain. Graph properties, which are set specifically for each graph run (i.e. RUN_ID), are overriddenby last graph.

For now, there are these types of graph events:

• graph started (p. 55)

• graph phase finished (p. 55)

• graph finished OK (p. 55)

• graph error (p. 55)

• graph aborted (p. 56)

• graph timeout (p. 56)

• graph status unknown (p. 56)

graph started

Event of this type is created, when ETL graph execution successfully started.

graph phase finished

Event of this type is created, everytime when graph phase is finished and all its nodes are finished with statusFINISHED_OK.

graph finished OK

Event of this type is created, when all phases and nodes of graph are finished with status FINISHED_OK.

graph error

Event of this type is created, when graph cannot be executed from any reason, or when any node of graph fails.

Chapter 7. Graph Event Listeners

56

graph aborted

Event of this type is created, when graph is explicitly aborted.

graph timeout

Event of this type is created, when graph runs longer then specified interval. Thus you have to specify "Job timeoutinterval" attribute for each listener of graph timeout event. You can specify this interval in seconds or in minutesor in hours.

Figure 7.1. Web GUI - graph timeout event

graph status unknown

Event of this type is created, when the server, during the startup, detects run records with undefined status in theexecutions history. Undefined status means, that server has been killed during graph run. Server automaticallychanges state of graph to "Not Available" and sends 'graph status unknown' event. Please note, that this worksjust for executions, which have persistent record in executions history. It is possible to execute transformationwithout persistent record in executions history, typically for better performance of fast running transformations(i.e. using Launch Services).

Listener

User may create listener for specified event type and graph (or all graphs in sandbox). Listener is actuallyconnection between graph event and task, where graph event specifies WHEN and task specifies WHAT to do.

So progress is like this:

• event is created

• listeners for this event are notified

• each listener performs related task

Tasks

Task types "execute shell command", "execute graph" and "archivator" are described in section "scheduling" seethis section for details about these task types. There is one more task type, which is useful especially with graphevent listeners, thus it is described here. It is task type "send email".

Note: You can use task of any type for both scheduling and graph event listener. Description of task types is dividedinto two sections just to show the most obvious use cases.


57

• Task - Send Email (p. 57)

• Task - JMS Message (p. 59)

Task - Send Email

This type of task is useful for notifications about result of graph execution. I.e. you can create listener with thistask type to be notified about each failure in specified sandbox or failure of particular graph.

Table 7.1. Attributes of "Send email" task

Task type "email"

Email pattern This select box contains all predefined email patterns. If user chooses anyof them, all fields below are automatically filled by values from pattern.

To Recipient's email address. It is possible to specify more addressesseparated by comma. It is also possible to use placeholders. SeePlaceholders (p. 58) for details.

Cc Cc stands for 'carbon copy'. Copy of email will be delivered to theseaddresses. It is possible to specify more addresses separated by comma. Itis also possible to use placeholders. See Placeholders (p. 58) for details.

BCc Bcc: stands for 'Blind carbon copy'. It is the same as Cc, but the othersrecipients aren't aware, that these recipients get copy of email.

Reply-to (Sender) Email address of sender. It must be valid address according to SMTP server.It is also possible to use placeholders. See Placeholders (p. 58) fordetails.

Subject Email subject. It is also possible to use placeholders. SeePlaceholders (p. 58) for details.

Plain text Body of email in plain text. Email is created as multipart, so HTMLbody should have a precedence. Plain text body is only for email clientswhich do not display HTML. It is also possible to use placeholders. SeePlaceholders (p. 58) for details.

HTML Body of email in HTML. Email is created as multipart, so HTML bodyshould have a precedence. Plain text body is only for email clientswhich do not display HTML. It is also possible to use placeholders. SeePlaceholders (p. 58) for details.

Log file as attachment If this switch is checked, email will have an attachment with packed logfile of related graph execution.


58

Figure 7.2. Web GUI - send email

Note: Do not forget to configure connection to SMTP server (See Chapter 18, Configuration (p. 96) for details).

Placeholders

Place holder may be used in some fields of tasks. They are especially useful for email tasks, where you can generatecontent of email according to context variables.

Note: In most cases, you can avoid this by using email patterns (See Email task for details)

These fields are preprocessed by Apache Velocity templating engine. See Velocity project URL for syntaxdescription http://velocity.apache.org/

There are several context variables, which you can use in place holders and even for creating loops and conditions.

• event

• now

• user

• run

• sandbox

Some of them may be empty in dependence of occasion which field is processed in. I.e. If task is processed becauseof graph event, then run and sandbox variables contain related data, otherwise they are empty,

http://velocity.apache.org/


59

Table 7.2. Placeholders useful in email templates

Variable name Contains

now Current date-time

user User, who caused this event. It may be owner of schedule, or someone whoexecuted graph. Contains sub-properties, which are accessible using dot notation (i.e.${user.email}) email, username, firstName, lastName, groups (list of values)

run Data structure describing one single graph execution. Contains sub-properties,which are accessible using dot notation (i.e. ${run.graphId}) graphId, finalStatus,startTime, stopTime, errNode, errMessage, errException, logLocation

tracking Data structure describing status of components in graph execution. Contains sub-properties, which are accessible using Velocity syntax for loops and conditions.

#if (${tracking})<table border="1" cellpadding="2" cellspacing="0">#foreach ($phase in $tracking.trackingPhases)<tr><td>phase: ${phase.phaseNumber}</td> <td>${phase.execTime} ms</td> <td></td><td></td><td></td></tr> #foreach ($node in $phase.trackingNodes) <tr><td>${node.nodeName}</td> <td>${node.result}</td> <td></td><td></td><td></td></tr> #foreach ($port in $node.trackingPorts) <tr><td></td><td></td> <td>${port.portType}:${port.index}</td> <td>${port.totalBytes} B</td> <td>${port.totalRows} rows</td></tr> #end #end#end</table>#end}

sandbox Data structure describing sandbox containing executed graph. Contains sub-properties, which are accessible using dot notation (i.e. ${sandbox.name}) name,code, rootPath

schedule Data structure describing schedule which triggered this task. Contains sub-properties, which are accessible using dot notation (i.e. ${schedule.description})description, startTime, endTime, lastEvent, nextEvent, fireMisfired

Task - JMS Message

This type of task is useful for notifications about result of graph execution. I.e. you can create graph event listenerwith this task type to be notified about each failure in specified sandbox or failure of particular graph.

JMS messaging requires JMS API (jms.jar) and third-party libraries. All these libraries must be available onapplication server classpath. Some application servers contain these libraries by default, some do not, thus thelibraries must be added explicitly.


60

Table 7.3. Attributes of JMS message task

Task type "JMS message"

Initial context class name Full class name of javax.naming.InitialContext implementation. EachJMS provider has own implementation. i.e. for Apache MQ it is"org.apache.activemq.jndi.ActiveMQInitialContextFactory". If it is empty,server uses default initial context

Connection factory JNDI name JNDI name of connection factory. Depends on JMS provider.

Destination JNDI name of message queue/topic on the server

Username Username for connection to JMS message broker

Password Password for connection to JMS message broker

URL URL of JMS message broker

JMS pattern This select box contains all predefined JMS message patterns. If userchooses any of them, text field below is automatically filled by value frompattern.

Text Body of JMS message. It is also possible to use placeholders. SeePlaceholders (p. 58) of send email task for details.

Figure 7.3. Web GUI - Task JMS message editor

Use cases

Possible use cases are the following:

• Execute graphs in chain (p. 61)

• Email notification about graph failure (p. 61)

• Email notification about graph success (p. 62)

• Backup of data processed by graph (p. 62)


61

Execute graphs in chain

Let's say, that we have to execute graph B, only if another graph A finished without any error. So there is somekind of relation between these graphs. We can achieve this behaviour by creating graph event listener. We createlistener for event graph finished OK of graph A and choose task type execute graph with graph Bspecified for execution. And that is it. If we create another listener for graph B with task execute graph withgraph C specified, it will work as chain of graphs.

Figure 7.4. Event source graph isn't specified, thus listener works for all graphs in specified sandbox

Email notification about graph failure

Figure 7.5. Web GUI - email notification about graph failure


62

Email notification about graph success

Figure 7.6. Web GUI - email notification about graph success

Backup of data processed by graph

Figure 7.7. Web GUI - backup of data processed by graph

63

Chapter 8. Jobflow Event ListenersPlease note that Jobflow Event Listeners work very similar to Graph Event Listener (the section called“Tasks” (p. 56)) in many ways, since ETL Graph and Jobflow are both "jobs" from the point of view of CloverETLServer.

In cluster environment, event exists only on cluster node, which runs jobflow, thus if the node isn't explicitlyspecified, the task is triggered on the same node.

Jobflow Events

Each event carries properties of the event source job. If there is a event listener specified, task may use theseproperties. e.g. next job in the chain may use "EVENT_FILE_NAME" placeholder which activated first job in thechain. Job properties, which are set specifically for each run (e.g. RUN_ID), are overridden by last job.

There are these types of jobflow events:

• jobflow started (p. 63)

• jobflow phase finished (p. 63)

• jobflow finished OK (p. 63)

• jobflow error (p. 63)

• jobflow aborted (p. 63)

• jobflow timeout (p. 64)

• jobflow status unknown (p. 64)

jobflow started

Event of this type is created, when jobflow execution successfully started.

jobflow phase finished

Event of this type is created, everytime when jobflow phase is finished and all its nodes are finished with statusFINISHED_OK.

jobflow finished OK

Event of this type is created, when all phases and nodes of jobflow are finished with status FINISHED_OK.

jobflow error

Event of this type is created, when jobflow cannot be executed from any reason, or when any node of the jobflowfails.

jobflow aborted

Event of this type is created, when jobflow is explicitly aborted.

Chapter 8. Jobflow Event Listeners

64

jobflow timeout

Event of this type is created, when jobflow runs longer then specified interval. Thus you have to specify "Jobtimeout interval" attribute for each listener of jobflow timeout event. You can specify this interval in seconds orin minutes or in hours.

Figure 8.1. Web GUI - jobflow timeout event

jobflow status unknown

Event of this type is created, when the server, during the startup, detects run records with undefined status in theexecutions history. Undefined status means, that server has been killed during jobflow run. Server automaticallychanges state of jobflow to "Not Available" and sends 'jobflow status unknown' event. Please note, that this worksjust for executions, which have persistent record in executions history. It is possible to execute transformationwithout persistent record in executions history, typically for better performance of fast running transformations(e.g. using Launch Services).

Listener

User may create listener for specified event type and jobflow (or all jobflows in sandbox). Listener is actuallyconnection between jobflow event and task, where jobflow event specifies WHEN and task specifies WHAT to do.

So progress is like this:

• event is created

• listeners for this event are notified

• each listener performs related task

Tasks

Task specifies operation which should be performed as the reaction to the triggered event.

Task types are described in the section called “Tasks” (p. 48) and the section called “Tasks” (p. 56)

Note: You can use task of any type for jobflow event listener. Description of task types is divided into two sectionsjust to show the most obvious use cases.

65

Chapter 9. JMS messages listenersThis feature allows you to specify listener for incoming JMS messages. Such listener can then process one ofpredefined tasks as usual for all event listeners. So for each listener user specifies source of JMS messages (JMSTopic or JMS Queue) and task which will be processed as a result of each incoming JMS message.

JMS messaging requires JMS API (jms.jar) and third-party libraries. All these libraries must be available onapplication server classpath. Some application servers contain these libraries by default, some do not, thus thelibraries must be added explicitly.

JMS itself is quite complex topic beyond of scope of this document. Detail information about it can be found onSun web site: http://java.sun.com/j2ee/1.4/docs/tutorial/doc/JMS6.html

Table 9.1. Attributes of JMS message task


Node ID to handle theevent

This attribute makes sense only in cluster environment. It is node ID where thelistener should be initialized. If it is not set, listener is initialized on all nodes in thecluster.

Initial context class name Full class name of javax.naming.InitialContext implementation. EachJMS provider has own implementation. i.e. for Apache MQ it is"org.apache.activemq.jndi.ActiveMQInitialContextFactory". If it is empty, serveruses default initial context. Specified class must be on web-app classpath orapplication-server classpath. It is usually included in one library with JMS APIimplementation for each specific JMS broker provider.

Connection factory JNDIname

JNDI name of connection factory. Depends on JMS provider.

Destination JNDI name JNDI name of message queue/topic on the server

Username Username for connection to JMS message broker

Password Password for connection to JMS message broker

URL URL of JMS message broker

Durable subscriber (onlyfor Topics)

If it is false, message consumer is connected to the broker as "non-durable", so itreceives only messages which are sent while the connection is active. Other messagesare lost. If it is true, consumer is subscribed as "durable" so it receives even messageswhich are sent while the connection is inactive. The broker stores such messagesuntil they can be delivered or until the expiration is reached. This switch makes senseonly for Topics destinations, because Queue destinations always store messages untilthey can be delivered or the expiration is reached. Please note, that consumer isinactive i.e. during server restart and during short moment when user updates the"JMS message listener" ant it must be re-initialized. So during these intervals themessage in the Topic may get lost if the consumer does not have durable subscription.

If the subscription is durable, client must have "ClientId" specified. This attributecan be set in different ways in dependence of JMS provider. I.e. for ActiveMQ, it isset as URL parameter tcp://localhost:1244?jms.clientID=TestClientID

Message selector This "query string" can be used as specification of conditions for filtering incomingmessages. Syntax is well described on Java EE API web site: http://java.sun.com/j2ee/1.4/docs/api/javax/jms/Message.html It has different behaviour depending ontype of consumer (queue/topic) Queue: If a its a queue the messages that are filteredout remain on the queue. Topic: Messages filtered out by a Topic subscriber'smessage selector will never be delivered to the subscriber. From the subscriber'sperspective, they do not exist.

Groovy code Groovy code may be used for additional message processing and/or for refusingmessage. Both features are described below.

http://java.sun.com/j2ee/1.4/docs/tutorial/doc/JMS6.html

http://java.sun.com/j2ee/1.4/docs/api/javax/jms/Message.html

http://java.sun.com/j2ee/1.4/docs/api/javax/jms/Message.html

Chapter 9. JMS messages listeners

66

Optional Groovy code

Groovy code may be used for additional message processing or for refusing message.

• Additional message processing Groovy code may modify/add/remove values stored in containers "properties"and "data".

• Refuse/acknowledge the message if Groovy code returns Boolean.FALSE, message is refused. Otherwise,message is acknowledged. Refused message may be redelivered, however JMS broker should have configuredsome limit for redelivering messages. If groovy code throws an exception, it is considered as coding error andJMS message is NOT refused because of it. So if the message refusal is directed by some exception, it mustbe handled in groovy.

Table 9.2. Variables accessible in groovy code

type key description

javax.jms.Message msg instance of JMS message

java.util.Properties properties See below for details. Contains values (Stringor converted to String) read from message andit is passed to the task which may use themsomehow. I.e. task "execute graph" passes theseparameters to the executed graph.

java.util.Map<String, Object> data See below for details. Contains values (Object,Stream, ..) read or proxied from the messageinstance and it is passed to the task which mayuse them somehow. I.e. task "execute graph"passes it to the executed graph as "dictionaryentries".

javax.servlet.ServletContext servletContext instance of ServletContext

javax.jms.Message msg instance of JMS message

com.cloveretl.server.api.ServerFacade serverFacade instance of serverFacade usable for callingCloverETL Server core features.

String sessionToken sessionToken, needed for calling serverFacademethods

Message data available for further processing

JMS message is processed and data it contains is stored basically in two data structures. "properties" and "data"


67

Table 9.3. "properties" elements

key description

JMS_PROP_[propertykey]

For each message property is created one entry, where "key" is made of prefix"JMS_PROP_" and property key.

JMS_MAP_[map entrykey]

If the message is instance of MapMessage, for each map entry is created oneentry, where "key" is made of prefix "JMS_MAP_" and map entry key. Values areconverted to String.

JMS_TEXT If the message is instanceof TextMessage, this property contains content of themessage.

JMS_MSG_CLASS Class name of message implementation

JMS_MSG_CORRELATIONIDCorrelation ID is either provider-specific message ID or application-specific Stringvalue

JMS_MSG_DESTINATIONThe JMSDestination header field contains the destination to which the message isbeing sent.

JMS_MSG_MESSAGEID A JMSMessageID is a String value that should function as a unique key foridentifying messages in a historical repository. The exact scope of uniqueness isprovider-defined. It should at least cover all messages for a specific installation of aprovider, where an installation is some connected set of message routers.

JMS_MSG_REPLYTO Destination to which a reply to this message should be sent.

JMS_MSG_TYPE Message type identifier supplied by the client when the message was sent.

JMS_MSG_DELIVERYMODEThe DeliveryMode value specified for this message.

JMS_MSG_EXPIRATION The time the message expires, which is the sum of the time-to-live value specifiedby the client and the GMT at the time of the send.

JMS_MSG_PRIORITY The JMS API defines ten levels of priority value, with 0 as the lowest priority and9 as the highest. In addition, clients should consider priorities 0-4 as gradations ofnormal priority and priorities 5-9 as gradations of expedited priority.

JMS_MSG_REDELIVERED"true" if this message is being redelivered.

JMS_MSG_TIMESTAMP The time a message was handed off to a provider to be sent. It is not the time themessage was actually transmitted, because the actual send may occur later due totransactions or other client-side queueing of messages.

Please note, that all values in "properties" structure are of String type, nevertheless it is number or text.

All listed properties are accessible with lower-case keys as well for backwards compatibility, however it'sdeprecated approach.

Table 9.4. "data" elements

key description

JMS_MSG instance of javax.jms.Message

JMS_DATA_STREAM Instance of java.io.InputStream. Accessible only for TextMessage, BytesMessage,StreamMessage, ObjectMessage(only if payload object is instance of String). Stringsare encoded in UTF-8.

JMS_DATA_TEXT Instance of String. Only for TextMessage and ObjectMessage, where payload objectis instance of String.

JMS_DATA_OBJECT Instance of java.lang.Object - message payload. Only for ObjectMessage.

"data" container is passed to the task which may use them somehow according to its implementation. I.e. task"execute graph" passes it to the executed graph as "dictionary entries". Please note that it is not serializable, thusif the task is relying on it, it can be processed properly only on the same cluster node.


68

Dictionary entries can be used in some of graph component attributes. I.e. in fileURL attribute like this:"dict:JMS_DATA_STREAM:discrete". So the reader reads data directly from incoming JMS message using thisproxy stream.

All listed dictionary entries are accessible with lower-case keys as well for backwards compatibility, however it'sdeprecated approach.

69

Chapter 10. Universal event listenersSince 2.10

This feature allows you to specify Groovy code, which decides when the event is created. Subsequently specifiedtask is processed. So for each listener user specifies Groovy source code and task which will be processed if groovycode decides to.

Table 10.1. Attributes of Universal message task


Node ID to handle theevent

This attribute makes sense only in cluster environment. It is node ID where thelistener should be initialized. If it is not set, listener is initialized on all nodes in thecluster.

Interval of check inseconds

Periodicity of Groovy code execution.

Groovy code Groovy code is used for deciding whether the event should be created or not. Seebelow for details about groovy code.

Groovy code

Groovy code is used for deciding whether the event should be created or not.

i.e. it may do some checks of data sources, which are vital for execution of graph. Or it may do some complexchecks of running graph and make decision to kill it. It may call CloverETL Server core functions usingServerFacade interface, which is described in its own chapter.

Creating "event" is simple. If Groovy code returns Boolean.TRUE, event is created and specified task is processed.Otherwise, nothing happens. If groovy code throws an exception, it is considered as coding error and event is NOTcreated because of it. So if it is necessary, the exceptions must be handled in groovy code.



java.util.Properties properties Empty container which may be filled by String-String key-value pairs in your Groovy code.It is passed to the task which may use themsomehow. I.e. task "execute graph" passes theseparameters to the executed graph.

java.util.Map<String, Object> data Empty container which may be filled by String-Object key-value pairs in your Groovy code.It is passed to the task which may use themsomehow according to its implementation. I.e.task "execute graph" passes it to the executedgraph as "dictionary entries". Please note that itis not serializable, thus if the task is relying onit, it can be processed properly only on the samecluster node.


com.cloveretl.server.api.ServerFacade serverFacade instance of serverFacade usable for callingCloverETL Server core features.


70

Chapter 11. Manual task executionSince 3.1

Manual task execution allows user to invoke task processing. Task is entity which describes how to react to somesource event. So normally task is processed only as a response to some source event. Since 3.1 user can manuallyinvoke task processing.

In addition user can specify some parameters to simulate source event which would normally trigger taskprocessing. Following figure displays how could be simulated "file event". Parameters for various event sourcesare listed in section "Graph parameters"

Figure 11.1. Web GUI - "Manual task execution" section

71

Chapter 12. File event listenersSince 1.3

File event listener allows system to monitor changes on server filesystem. User may define, which filesystemresource should be observed as a source of file event. User also specifies task, which should be processed asreaction to change on filesystem.

There is process which performs checks for changes on file system. This process works with preconfiguredperiodicity, thus there is minimal interval which for checks. You can set this minimal interval by clover property"clover.event.fileCheckMinInterval".

In cluster environment, each event listener has attribute "node ID" which specifies cluster node, which checks itslocal filesystem. In "standalone" environment, "node ID" attribute is ignored.

Figure 12.1. Web GUI - "File event listeners" section

Observed fileObserved file is specified by directory path and file name pattern.

User may specify just one exact file name or file name pattern for observing more matching files in specifieddirectory. If there are more changed files matching the pattern, separated event is triggered for each of these files.

There are three ways how to specify file name pattern of observed file(s)

• Exact match (p. 71)

• Wildcards (p. 71)

• Regullar expression (p. 72)

Exact match

You specify exact name of the observed file.

Wildcards

You can use wildcards common in most operating systems (*, ?, etc.)

Chapter 12. File event listeners

72

• * - Matches zero or more instances of any character

• ? - Matches one instance of any character

• [...] - Matches any of characters enclosed by the brackets

• \ - Escape character

Examples

• *.csv - Matches all CSV files

• input_*.csv - Matches i.e. input_001.csv, input_9.csv

• input_???.csv - Matches i.e. input_001.csv, but does not match input_9.csv

Regullar expression

Examples

• (.*?)\.(jpg|jpeg|png|gif)$ - Matches image files

Notes

• It is strongly recommended to use absolute paths. It is possible to use relative path, but working directorydepends on application server.

• Use forward slashes as file separators, even on MS Windows OS. Backslashes might be evaluated as escapesequences.

File Events

For each listener you have to specify event type, which you are interested in.

There are four types of file events:

• file APPEARANCE (p. 72)

• file DISAPPEARANCE (p. 72)

• file SIZE (p. 73)

• file CHANGE_TIME (p. 73)

file APPEARANCE

Event of this type occurs, when the observed file is created or copied from another location between two checks.Please keep in mind, that event of this type occurs immediately when new file is detected, regardless it is completeor not. Thus task which may need complete file is executed when file is still incomplete. Recommended approach isto save file to the different location and when it is complete, move/rename to observed location where CloverETLServer may detect it. File moving/renaming should be atomic operation.

Event of this type does not occur when the file has been updated (change of timestamp or size) between twochecks. Appearance means that the file didn't exist during previous check and it exists now, during current check.

file DISAPPEARANCE

Event of this type occurs, when observed file is deleted or moved to another location between two checks.

Chapter 12. File event listeners

73

file SIZE

Event of this type occurs when the size of the observed file has changed between two checks. Event of this typeis never produced when file is created or removed. File must exist during both checks.

file CHANGE_TIME

Event of this type occurs, when change time of observed file has changed between two checks. Event of this typeis never produced when file is created or removed. File must exist during both checks.

Check interval, Task and Use cases

• User may specify minimal time interval between two checks. It is specified in seconds.

• Each listener defines task, which will be processed as the reaction for file event. All task types and theirsattributes are described in section Scheduling and GraphEventListeners

• • Graph Execution, when file with input data is accessible

• Graph Execution, when file with input data is updated

• Graph Execution, when file with generated data is removed and must be recreated

How to use source of event during task processing

File, which caused event (considered as source of event) may be used during task processing. i.e. reader/writercomponents in graph transformations may refer to this file by special placeholders: ${EVENT_FILE_PATH} -path to directory which contains event source ${EVENT_FILE_NAME} - name of event source.

Please note that previous versions used lower-case placeholders. Since version 3.3, placeholders are upper-case,however lower-case still work for backward compatibility.

i.e. if event source is: /home/clover/data/customers.csv, placeholders will contain:EVENT_FILE_PATH - /home/clover/data, EVENT_FILE_NAME - customers.csv

For "graph execution" task this works only if the graph is not pooled. Thus "keep in pool interval" must be setto 0 (default value).

74

Chapter 13. WebDAVSince 3.1

WebDAV API allows user to use standard WebDAV clients for managing sandboxes content.

It allows specifically:

• browsing directory structure

• editing files

• removing files/folders

• renaming files/folders

• creating files/folders

• copying files

• moving files

It is accessible on URL "http://[host]:[port]/clover/webdav".

Although common www browsers can open this URL, most of them are not rich WebDAV clients, thus you canjust see list of items, but you cannot browse the directory structure with common www browsers.

WebDAV clients

There are many WebDAV clients for various operating systems, some OS support WebDAV natively.

Linux like OS

Great WebDAV client working on linux systems is Konqueror. Please use different protocol in the URL: webdav://[host]:[port]/clover/webdav

MS windows

Last distributions of MS Windows (Win XP and later) have native support for WebDAV. Unfortunatelly, it ismore or less unreliable, so it is recommended to use some free or commercial WebDAV client.

• The best WebDAV client we've tested is BitKinex: http://www.bitkinex.com/webdavclient

• Another option is to use Total Commander (http://www.ghisler.com/index.htm) with WebDAV plugin: http://www.ghisler.com/plugins.htm#filesys

Mac OS

Mac OS supports WebDAV natively and in this case it should be without any problems. You can use "finder"application, select "Connect to the server ..." menu item and use URL with HTTP protocol: "http://[host]:[port]/clover/webdav".

WebDAV authentication/authorization

CloverETL Server WebDAV API uses the HTTP Basic Authentication by default. However it may be reconfiguredto use HTTP Digest Authentication. Please see the Configuration section for details.

Chapter 13. WebDAV

75

Digest Authentication may be useful, since some WebDAV clients can't work with HTTP Basic Authentication,only with Digest Authentication.

HTTP Digest Authentication is feature added to the version 3.1. If you upgraded your older CloverETL Serverdistribution, users created before the upgrade cannot use the HTTP Digest Authentication until they reset theirpasswords. So when they reset their passwords (or the admin does it for them), they can use Digest Authenticationas well as new users.

76

Chapter 14. Simple HTTP APIThis API is intended for all HTTP clients (even for the simplest ones - like wget tool). All operations are accessibleusing http GET method and return plain text. Thus response can be parsed by simple tools. If global security is on(default setting), BASIC HTTP authentication is required. Use CloverETL Server user with proper permissions.

Please note, that ETL graph related operations "graph_run", "graph_status" and "graph_kill" work for jobflowsas well.

URL has this pattern:

http://[domain]:[port]/[context]/[servlet]/[operation]?[param1]=[value1]&[param2]=[value2]...

For wget client, you can use following command line:

wget --user=$USER --password=$PASS -O ./$OUTPUT_FILE $REQUEST_URL

• Operation help (p. 76)

• Operation graph_run (p. 77)

• Operation graph_status (p. 77)

• Operation graph_kill (p. 78)

• Operation server_jobs (p. 79)

• Operation sandbox_list (p. 79)

• Operation sandbox_content (p. 79)

• Operation executions_history (p. 79)

• Operation suspend (p. 81)

• Operation resume (p. 81)

• Operation sandbox_create (p. 82)

• Operation sandbox_add_location (p. 82)

• Operation sandbox_remove_location (p. 82)

• Cluster status (p. 83)

Operation help

parameters

no

returns

list of possible operations and parameters with its descriptions

Chapter 14. Simple HTTP API

77

example

http://localhost:8080/clover/request_processor/help

Operation graph_run

Call this operation to start execution of the specified job. Operation is called graph_run for backward compatibility,however it may execute ETL graph or jobflow.

parameters

Table 14.1. Parameters of graph_run

parameter name mandatory default description

graphID yes - Text Id, which is unique in specified sandbox. May be filepath relative to sandbox root

sandbox yes - Text ID of sandbox

additional graphparameters

no Any URL parameter with "param_" prefix is passedto executed graph and may be used in graph XMLas placeholder, but without "param_" prefix. i.e."param_FILE_NAME" specified in URL may be used in thegraph as ${FILE_NAME}. These parameters are resolvedonly during loading of graph XML, so graph cannot bepooled.

nodeID no - In cluster mode it's ID of node which should execute the job.However it's not final. If the graph is distributed, or the nodeis disconnected, the graph may be executed on some anothernode.

verbose no MESSAGE MESSAGE | FULL - how verbose should possible errormessage be.

returns

run ID: incremental number, which identifies each execution request

example

http://localhost:8080/clover/request_processor/graph_run?graphID=graph/graphDBExecute.grf&sandbox=mva

Operation graph_status

Call this operation to obtain status of specified job execution. Operation is called graph_status for backwardcompatibility, however it may return status of ETL graph or jobflow.

parameters


78

Table 14.2. Parameters of graph_status


runID yes - Id of each graph execution

returnType no STATUS STATUS | STATUS_TEXT | DESCRIPTION |DESCRIPTION_XML

waitForStatus no - Status code which we want to wait for. If it is specified, thisoperation will wait until graph is in required status.

waitTimeout no 0 If waitForStatus is specified, it will wait only specifiedamount of milliseconds. Default 0 means forever, but itdepends on application server configuration. When thespecified timeout expires and graph run still isn't in requiredstatus, server returns code 408 (Request Timeout). 408 codemay be also returned by application server if its HTTPrequest timeout expires before.


returns

Status of specified graph. It may be number code, text code or complex description in dependence of optionalparameter returnType. Description is returned as plain text with pipe as separator, or as XML. Schema describingXML format of the XML response is accessible on CloverETL Server URL: http://[host]:[port]/clover/schemas/executions.xsd In dependence on waitForStatus parameter may return result immediately or wait for specifiedstatus.

example

http://localhost:8080/clover/request_processor/graph_status -> -> ?runID=123456&returnType=DESCRIPTION&waitForStatus=FINISHED&waitTimeout=60000

Operation graph_kill

Call this operation to abort/kill job execution. Operation is called graph_kill for backward compatibility, howeverit may abort/kill ETL graph or jobflow.

parameters

Table 14.3. Parameters of graph_kill


runID yes - Id of each graph execution

returnType no STATUS STATUS | STATUS_TEXT | DESCRIPTION


returns

Status of specified graph after attempt to kill it. It may be number code, text code or complex description independence of optional parameter.

example

http://localhost:8080/clover/request_processor/graph_kill?runID=123456&returnType=DESCRIPTION


79

Operation server_jobs

parameters

no

returns

List of runIDs currently running jobs.

example

http://localhost:8080/clover/request_processor/server_jobs

Operation sandbox_list

parameters

no

returns

List of all sandbox text IDs. In next versions will return only accessible ones.

example

http://localhost:8080/clover/request_processor/sandbox_list

Operation sandbox_content

parameters

Table 14.4. Parameters of sandbox_content


sandbox yes - text ID of sandbox


returns

List of all elements in specified sandbox. Each element may be specified as file path relative to sandbox root.

example

http://localhost:8080/clover/request_processor/sandbox_content?sandbox=mva

Operation executions_history

parameters


80

Table 14.5. Parameters of executions_history


sandbox yes - text ID of sandbox

from no Lower datetime limit. Operation will return only recordsafter(and equal) this datetime. Format: "yyyy-MM-ddHH:mm" (must be URL encoded)

to no Lower datetime limit. Operation will return only recordsafter(and equal) this datetime. Format: "yyyy-MM-ddHH:mm" (must be URL encoded) status

status no Current execution status. Operation will return only recordswith specified STATUS. Meaningful values are RUNNING| ABORTED | FINISHED_OK | ERROR

sandbox no Sandbox code. Operation will return only records for graphsfrom specified sandbox.

graphId no Text Id, which is unique in specified sandbox. File pathrelative to sandbox root

orderBy no Attribute for list ordering. Possible values: id | graphId |finalStatus | startTime | stopTime. There is no ordering bydefault.

orderDescend no true Switch which specifies ascending or descending ordering. Ifit is true (which is default), ordering is descending.

returnType no IDs Possible values are: IDs | DESCRIPTION |DESCRIPTION_XML

index no 0 Index of the first returned records in whole record set.(starting from

records no infinite Max amount of returned records.


returns

List of executions according to filter criteria.

For returnType==IDs returns simple list of runIDs (with new line delimiter).

For returnType==DESCRIPTION returns complex response which describes current status of selectedexecutions, their phases, nodes and ports.

execution|[runID]|[status]|[username]|[sandbox]|[graphID]|[startedDatetime]|[finishedDatetime]|[clusterNode]|[graphVersion]phase|[index]|[execTimeInMilis]node|[nodeID]|[status]|[totalCpuTime]|[totalUserTime]|[cpuUsage]|[peakCpuUsage]|[userUsage]|[peakUserUsage]port|[portType]|[index]|[avgBytes]|[avgRows]|[peakBytes]|[peakRows]|[totalBytes]|[totalRows]

example of request

http://localhost:8080/clover/request_processor/executions_history -> -> ?from=&to=2008-09-16+16%3A40&status=&sandbox=def&graphID=&index=&records=&returnType=DESCRIPTION

example of DESCRIPTION (plain text) response

execution|13108|FINISHED_OK|clover|def|test.grf|2008-09-16 11:11:19|2008-09-16 11:11:58|nodeA|2.4


81

phase|0|38733node|DATA_GENERATOR1|FINISHED_OK|0|0|0.0|0.0|0.0|0.0port|Output|0|0|0|0|0|130|10node|TRASH0|FINISHED_OK|0|0|0.0|0.0|0.0|0.0port|Input|0|0|0|5|0|130|10node|SPEED_LIMITER0|FINISHED_OK|0|0|0.0|0.0|0.0|0.0port|Input|0|0|0|0|0|130|10port|Output|0|0|0|5|0|130|10execution|13107|ABORTED|clover|def|test.grf|2008-09-16 11:11:19|2008-09-16 11:11:30phase|0|11133node|DATA_GENERATOR1|FINISHED_OK|0|0|0.0|0.0|0.0|0.0port|Output|0|0|0|0|0|130|10node|TRASH0|RUNNING|0|0|0.0|0.0|0.0|0.0port|Input|0|5|0|5|0|52|4node|SPEED_LIMITER0|RUNNING|0|0|0.0|0.0|0.0|0.0port|Input|0|0|0|0|0|130|10port|Output|0|5|0|5|0|52|4

For returnType==DESCRIPTION_XML returns complex data structure describing one or more selectedexecutions in XML format. Schema describing XML format of the XML response is accessible on CloverETLServer URL: http://[host]:[port]/clover/schemas/executions.xsd

Operation suspend

Suspends server or sandbox(if specified). Suspension means, that no graphs may me executed on suspended server/sandbox.

parameters

Table 14.6. Parameters of suspend


sandbox no - Text ID of sandbox to suspend. If not specified, it suspendswhole server.

atonce no If this param is set to true, running graphs from suspendedserver(or just from sandbox) are aborted. Otherwise it canrun until it is finished in common way.

returns

Result message

Operation resume

parameters

Table 14.7. Parameters of resume


sandbox no - Text Id of sandbox to resume. If not specified, server willbe resumed.


returns

Result message


82

Operation sandbox_create

This operation creates specified sandbox. If it's sandbox of "partitioned" or "local" type, create also locations by"sandbox_add_location" operation.

parameters

Table 14.8. Parameters of sandbox create


sandbox yes - Text Id of sandbox to be created.

path no - Path to the sandbox root if server is running in standalonemode.

type no shared Sandbox type: shared | partitioned | local. For standaloneserver may be left empty, since the default "shared" is used.

createDirs no true Switch whether to create directory structure of the sandbox(only for standalone server or "shared" sandboxes in clusterenvironment).


returns

Result message

Operation sandbox_add_location

This operation adds location to specified sandbox. Only useable for sandboxes of type partitioned or local.

parameters

Table 14.9. Parameters of sandbox add location


sandbox yes - Sandbox which we want to add location to.

nodeId yes - Location attribute - node which has direct access to thelocation.

path yes - Location attribute - path to the location root on the specifiednode.

location no - Location attribute - location storage ID. If it's not specified,new one will be generated.


returns

Result message

Operation sandbox_remove_location

This operation removes location from specified sandbox. Only sandboxes of type partitioned or local can havelocations asociated.


83

parameters

Table 14.10. Parameters of sandbox add location


sandbox yes - Removes specified location from its sandbox.

location yes - Location storage ID. If the specified location isn't attachedto the specified sandbox, sandbox won't be changed.


returns

Result message

Cluster status

This operation displays cluster's nodes list.

parameters

no

returns

Cluster's nodes list.

84

Chapter 15. JMX mBeanCloverETL Server JMX mBean is API, which is useful for monitoring of CloverETL Server's internal status.

MBean is registered with name:

com.cloveretl.server.api.jmx:name=cloverServerJmxMBean

.

JMX configuration

Application's JMX MBeans aren't accessible outside of JVM by default. It needs some changes in applicationserver configuration to make them accessible.

This section describes how to configure JMX Connector for development and testing. Thus authentication may bedisabled. For production deployment authentication should be enabled. Please refer further documentation to seehow to achieve this. i.e. http://java.sun.com/j2se/1.5.0/docs/guide/management/agent.html#auth

Configurations and possible problems:

• How to configure JMX on Apache Tomcat (p. 84)

• How to configure JMX on Glassfish (p. 85)

• Websphere 7 (p. 86)

• Possible problems (p. 87)

How to configure JMX on Apache Tomcat

Tomcat's JVM must be executed with these self-explanatory parameters:

1. -Dcom.sun.management.jmxremote=true

2. -Dcom.sun.management.jmxremote.port=8686

3. -Dcom.sun.management.jmxremote.ssl=false

4. -Dcom.sun.management.jmxremote.authenticate=false

On UNIX like OS set environment variable CATALINA_OPTS i.e. like this:

export CATALINA_OPTS="-Dcom.sun.management.jmxremote=true -Dcom.sun.management.jmxremote.port=8686 -Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=false"

File TOMCAT_HOME/bin/setenv.sh (if it does not exist, you may create it) or TOMCAT_HOME/bin/catalina.sh

On Windows it might be tricky, that each parameter must be set separately:

set CATALINA_OPTS=-Dcom.sun.management.jmxremote=trueset CATALINA_OPTS=%CATALINA_OPTS% -Dcom.sun.management.jmxremote.port=8686set CATALINA_OPTS=%CATALINA_OPTS% -Dcom.sun.management.jmxremote.authenticate=false

http://java.sun.com/j2se/1.5.0/docs/guide/management/agent.html#auth

Chapter 15. JMX mBean

85

set CATALINA_OPTS=%CATALINA_OPTS% -Dcom.sun.management.jmxremote.ssl=false

File TOMCAT_HOME/bin/setenv.bat (if it does not exist, you may create it) or TOMCAT_HOME/bin/catalina.bat

With these values, you can use URL

service:jmx:rmi:///jndi/rmi://localhost:8686/jmxrmi

for connection to JMX server of JVM. No user/password is needed

How to configure JMX on Glassfish

Go to Glasfish admin console (by default accessible on http://localhost:4848 with admin/adminadmin as user/password)

Go to section "Configuration" > "Admin Service" > "system" and set attributes like this:

Figure 15.1. Glassfish JMX connector

With these values, you can use URL

service:jmx:rmi:///jndi/rmi://localhost:8686/jmxrmi

for connection to JMX server of JVM.

Use admin/adminadmin as user/password. (admin/adminadmin are default glassfish values)


86

How to configure JMX on Websphere

Websphere does not require any special configuration, but the clover MBean is registered with the name, thatdepends on application server configuration:

com.cloveretl.server.api.jmx:cell=[cellName],name=cloverServerJmxMBean,node=[nodeName], process=[instanceName]

Figure 15.2. Websphere configuration

Websphere 7

URL for connecting to JMX server is:

service:jmx:iiop://[host]:[port]/jndi/JMXConnector

where host is the host name you are connectiong to and port is RMI port number. If you have a default Websphereinstallation, the JNDI port number will likely be 2809, 2810, ... depending on how many servers there are installedon one system and the specific one you want to connect to. To be sure, when starting Websphere, check the logs,as it will dump a line like

0000000a RMIConnectorC A ADMC0026I: The RMI Connector is available at port 2810

How to configure JMX on Websphere7

Websphere does not require any special configuration, but the clover MBean is registered with the name, thatdepends on application server configuration:

com.cloveretl.server.api.jmx:cell=[cellName],name=cloverServerJmxMBean,node=[nodeName], process=[instanceName]


87

Figure 15.3. Websphere7 configuration

URL for connecting to JMX server is:

service:jmx:iiop://[host]:[port]/jndi/JMXConnector

where host is the host name you are connectiong to and port is RMI port number. If you have a default Websphereinstallation, the JNDI port number will likely be 2809, 2810, ... depending on how many servers there are installedon one system and the specific one you want to connect to. To be sure, when starting Websphere, check the logs,as it will dump a line like

0000000a RMIConnectorC A ADMC0026I: The RMI Connector is available at port 2810

You will also need to set on the classpath following jar files from Websphere home directory:

/runtimes/com.ibm.ws.admin.client_7.0.0.jar/runtimes/com.ibm.ws.ejb.thinclient_7.0.0.jar/runtimes/com.ibm.ws.orb_7.0.0.jar

Possible problems

• Default JMX mBean server uses RMI as a transport protocol. Sometimes RMI cannot connectremotely when one of peers uses Java version 1.6. Solution is quite easy, just set these twosystem properties: -Djava.rmi.server.hostname=[hostname or IP address] -Djava.net.preferIPv4Stack=true

Operations

For details about operations please see the JavaDoc of the MBean interface:

JMX API MBean JavaDoc is accessible in the running CloverETL Server instance on URL: http://[host]:[port]/[contextPath]/javadoc-jmx/index.html

../javadoc-jmx/index.html

../javadoc-jmx/index.html

88

Chapter 16. SOAP WebService APICloverETL Server SOAP Web Service is API, which allows its clients to manipulate with content of the sandboxes,to monitor status of executed graphs and more.

Service is accessible on URL:

http://[host]:[port]/clover/webservice

Service descriptor is accessible on URL:

http://[host]:[port]/clover/webservice?wsdl

Protocol HTTP can be changed to secured HTTPS according to web server configuration.

SOAP WS Client

Exposed service is implemented with the most common binding style "document/literal", which is widelysupported by libraries in various programming languages.

To create client for this API, only WSDL document (see the URL above) is needed together with somedevelopment tools according to your programming language and development environments.

JavaDoc of WebService interface with all related classes is accessible in the running CloverETL Server instanceon URL http://[host]:[port]/[contextPath]/javadoc-ws/index.html

If the web server has HTTPS connector configured, also the client must meet the security requirements accordingto web server configuration. i.e. client trust + key stores configured properly

SOAP WS API authentication/authorization

Since exposed service is stateless, authentication "sessionToken" has to be passed as parameter to each operation.Client can obtain authentication sessionToken by calling "login" operation.

../javadoc-ws/index.html

89

Chapter 17. Launch ServiceThe Launch Service provides users with convenient way of remotely executing the CloverETL graphs or Jobflowsvia a simple web-based interface.

The Launch Services can be used with any HTTP client. This allows for convenient control of the job executionwhich can be easily tied to external tools if necessary (requests can be sent from custom applications as well).

Launch Service Overview

The architecture of Launch Service is relatively simple and follows the basic design of multi-tiered applicationsutilizing the browser.

The Launch Service and underlaying CloverETL Server jobs may work as back-end for any front-end. It may beclient's web application or third party application which calls CloverETL Launch Service by HTTP request. ThusLaunch Services allow full customization of the outside appearance of the web - for example, it can be a simpleweb form which communicates with users in the terminology they are familiar with.

Figure 17.1. Launch Services and CloverETL Server as web application back-end

Deploying Graph in Launch Service

To enable users to access the specific job via Launch Service, several steps have to be taken:

1. The job has to be designed to allow its parameters to be passed via dictionary.

2. The job has to be configured in CloverETL Server in Launch Service section.

3. The form which will submit the data to Launch Service has to be written.

Chapter 17. Launch Service

90

Overall the deployment of the job to the Launch Service is not much more complex compared to the regular jobdevelopment process. In following chapters all the steps will be described in more detail alongside some basicexamples.

Designing the ETL graph/Jobflow for Launch Service

To use the jobs from Launch Service, the Launch Service requires the job to use dictionary when parameters haveto be passed to the job. Dictionary is a data storage associated with each run of the job in CloverETL. For moredetails about the dictionary see section "Dictionary" in CloverETL Designer docs.

To use the Dictionary from the Launch Service, the job author is required to specify the entries of the dictionary injob's XML source file. For more details about the Dictionary XML element see section "Dictionary" in CloverETLDesigner docs.

Apart from the use of the dictionary, the Launch Service does not impose any other restriction on the jobs it shouldrun. The jobs can therefore use all the facilities provided by the CloverETL engine.

Configuring the job in CloverETL Server web GUI

To notify the Launch Service about the jobs that will be available via its interface, the Launch Service has to beproperly configured via CloverETL Server GUI.

Launch Service uses launch configurations to store the details about how each job can be run. Each launchconfiguration contains full description of the job's parameters, how they are mapped to the parameters passed fromthe web interface and so on.

Each launch configuration is identified by its name, user and group restriction. Several configurations with thesame name can be created as long as they differ in their user or group restrictions.

Use restrictions can be used to launch different jobs for different users even though they use the same launchconfiguration (for example, the developers may want to use debug version of the job while the end customers willwant to use the production job). The user restriction can also be used to prohibit certain users from executing thelaunch configuration.

Similarly, the group restriction can be used to differentiate jobs based on the group membership of the user whichruns the launch configuration.

When the configuration is launched, the correct configuration is picked based on the configuration name, userspecification and group specification. If multiple configuration match the current user/group and configurationname, the most specific one is picked (the user name has higher priority than the group name).

Adding New Launch Configuration

New launch configurations can be added by clicking on New launch configuration link on the Launch Servicestab in CloverETL Server GUI:


91

Figure 17.2. Launch Services section

After the configuration has been created it will appear in the table on the left side among the other existingconfiguration. Before using the configuration user will have to add parameter mapping. To add parametermappings click on the detail link for the newly created configuration. The details will be displayed on the rightside of the window in a simple table:

Figure 17.3. Overview tab

The Overview tab shows the basic details about the launch configuration. These can be modified in the EditConfiguration tab:


92

Figure 17.4. Edit Configuration tab

Following fields can be modified:

• Name - is the name under which the configuration will be accessible from web.

• Description - the description of the configuration.

• Group - restricts the configuration to specific group of users.

• User - restricts the configuration to specified user.

• Sandbox - selects the CloverETL Sandbox in which the configuration will be launched.

• Job file - selects the job to run when the configuration is launched.

• Save run record - if checked, the details about the launch configuration will be visible in Execution Historyin the CloverETL Server GUI. If unchecked, the job executions will not be logged and will not be displayedin the Execution History.

• Display error message detail - if checked, detailed error messages will be displayed in case the launch fails. Ifunchecked, only simpler messages will be displayed to the user.

Finally, the tab Edit Parameters can be used to configure parameter mappings for the launch configuration. Themappings are required for the Launch Service to be able to correctly assign parameters values based on the valuessent in the launch request.


93

Figure 17.5. Creating new parameter

To add new parameter mapping click on the New property link. Each property required by the job has to be created(internal job properties do not need mappings).

Figure 17.6. Edit Parameters tab

Following fields are available for each property:

• Name - the name of the property in the job's dictionary.

• Request parameter - the name of the parameter as specified in the launch request generated by the request page.This name can be different than the name used in job's dictionary.

• Parameter required - if checked the parameter is mandatory and error will be reported if it is omitted.

• Pass to job - if checked the parameter will be also passed to job among the additional parameters as well asin the dictionary. In such case, the parameter can also be referenced as ${ParameterName} in the job's XMLfile. Since the additional parameters are resolved when the XML file is parsed, the job which use this methodcannot be pooled.

• Default value - is the default value which will be applied in case the parameter is omitted in the launch request.

To create the new mapping, click on the Create button after all the fields have been filled. After the mappingis created, it will be displayed in the list of existing mappings. It can be later edited or deleted by clicking onappropriate links.


94

Sending the Data to Launch Service

To launch the job which has been configured for use with Launch Service, the user has to send a launch request. Thelaunch request can be sent via HTTP GET or POST methods. A launch request is simply an URL which containsthe values of all parameters that should be passed to the job. The request URL is composed of several parts:

[Clover Context]/launch/[Configuration name]?[Parameters]

• [Clover Context] is the URL to the context in which the CloverETL is running. Usually this isthe full URL to CloverETL Server (for example, for CloverETL Demo Server this would be http://server-demo.cloveretl.com:8080/clover).

• [Configuration name] is the name of the launch configuration which has been specified when theconfiguration has been created. In our example, this would be set to NewMountains (distinction between upper-and lower-case is important).

• [Parameters] is the list of parameters the configuration requires in the format used for example by PHP.Therefore the parameter list is a list of name-value pairs separated by "&" character. Each name-value pairis specified as [name]=[value] where value has to be properly encoded according to RFC 1738 to make sureURL is valid.

Based on the above, the full URL of launch request for our example with mountains may be like this: http://server-demo.cloveretl.com:8080/clover/launch/NewMountains?heightMin=4000. In the request above, the valueof heightMin property is set to 4000.

Results of the Graph Execution

After the job's run terminates, the results are sent back to the HTTP client as a content of HTTP response. Theoutput is partially defined in the dictionary which is declared in the job's XML file. The dictionary can markselected parameters as output parameters. All the output parameters are sent to the user after the job executionis finished.

Depending on the number of output parameters, the following output is sent to the HTTP client:

• No output parameters - only summary page is returned. The format of the summary page cannot be customized.The page will contain details like when the job was started, when it finished, user name and so on.

• One output parameter - in this case the output is sent to the HTTP client with its content type defined by theproperty type in the dictionary.

• Multiple output parameters - in this case each output parameter is sent to the HTTP client as a part of multipartHTTP response. The content type of the response is either multipart/related or multipart/x-mixed-replacedepending on the HTTP client (the client detection is of course fully automatic). The multipart/related typeis used for browsers based on Microsoft Internet Explorer, the multipart/x-mixed-replace is sent to browsersbased on Gecko or Webkit.

Launch requests are recorded in the log files in directory specified by launch.log.dir property in CloverETLServer configuration. For each launch configuration one log file named [Configuration name]#[Launch ID].log iscreated. For each launch request this file will contain only one line with following tab-delimited fields:

If the property launch.log.dir is not specified, log files are created in temp directory[java.io.tmpdir]/cloverlog/launch. Where "java.io.tmpdir" is system property.

• Launch start time

• Launch end time

• Logged-in user name


95

• Run ID

• Execution status FINISHED_OK, ERROR or ABORTED

• IP Address of the client

• User agent of the HTTP client

• Query string passed to the Launch Service (full list of parameters of the current launch)

In case the configuration is not valid, the same launch details are saved into the _no_launch_config.log file in thesame directory. All unauthenticated requests are saved to the same file as well.

96

Chapter 18. ConfigurationDefault installation (without any configuration) is recommended only for evaluation purposes. For production, atleast DB connection and SMTP server configuration is recommended.

Config Sources and Their Priorities

There are several sources of configuration properties. If property isn't set, application default is used.

Warning: Do not combine sources specified below. Configuration becomes confusing and maintenance will bemuch more difficult.

Context Parameters (Available on Apache Tomcat)

Some application servers allow to set context parameters without modification of WAR file. This way ofconfiguration is possible and recommended for Tomcat.

Example for Apache Tomcat

On Tomcat it is possible to specify context parameters in context configuration file. [tomcat_home]/conf/Catalina/localhost/clover.xml which is created automatically just after deployment of CloverETLServer web application.

You can specify property by adding this element:

<Parameter name="[propertyName]" value="[propertyValue]" override="false" />

Environment Variables

Set environment variable with prefix clover. , i.e. (clover.config.file)

Some operating systems may not use dot character, so also underlines (_) may be used instead of dots (.). So theclover_config_file works as well.

System Properties

Set system property with prefix clover. , i.e. (clover.config.file)

Also underlines (_) may be used instead of dots (.) so the clover_config_file works as well.

Properties File on default Location

Source is common properties file (text file with key-value pairs):

[property-key]=[property-value]

By default CloverETL tries to find config file [workingDir]/cloverServer.properties.

Properties File on specified Location

The same as above, but properties file is not loaded from default location, because its location is specifiedby environment variable or system property clover_config_file or clover.config.file. This isrecommended way of configuration if context parameters cannot be set in application server.

Chapter 18. Configuration

97

Modification of Context Parameters in web.xml

Unzip clover.war and modify file WEB-INF/web.xml, add this code:

<context-param> <param-name>[property-name]</param-name> <param-value>[property-value]</param-value></context-param>

This way isn't recommended, but it may be useful when none of above ways is possible.

Priorities of config Sources

Configuration sources have these priorities:

1. context parameters (specified in application server or directly in web.xml)

2. external config file CS tries to find it in this order (only one of them is loaded):

• path specified by context parameter config.file

• path specified by system property clover_config_file or clover.config.file

• path specified by environment variable clover_config_file or clover.config.file

• default location ([workingDir]/cloverServer.properties)

3. system properties

4. environment variables

5. default values

Examples of DB Connection ConfigurationConfiguration of DB connection is optional. Embedded Apache Derby DB is used by default and it is sufficientfor evaluation. However, configuration of external DB connection is strongly recommended for productiondeployment. It is possible to specify common JDBC DB connection attributes (URL, username, password) orJNDI location of DB DataSource.

Configurations and their changes may be as follows:

• Embedded Apache Derby (p. 97)

• MySQL (p. 98)

• DB2 (p. 98)

• Oracle (p. 100)

• MS SQL (p. 101)

• Postgre SQL (p. 101)

• JNDI DB DataSource (p. 101)

Embedded Apache Derby

Apache Derby embedded DB is used with default CloverETL Server installation. It uses working directory asstorage directory for data persistence by default. This may be problem on some systems. In case of any problems


98

with connecting to Derby DB, we recommend you configure connection to external DB or at least specify Derbyhome directory:

Set system property derby.system.home to set path which is accessible for application server. You canspecify this system property by this JVM execution parameter:

-Dderby.system.home=[derby_DB_files_root]

If you use properties file for configuration, specify these parameters: jdbc.driverClassName,jdbc.url, jdbc.username, jdbc.password, jdbc.dialect. For example:

jdbc.driverClassName=org.apache.derby.jdbc.EmbeddedDriverjdbc.url=jdbc:derby:databases/cloverDb;create=truejdbc.username=jdbc.password=jdbc.dialect=org.hibernate.dialect.DerbyDialect

Take a closer look at the jdbc.url parameter. Part databases/cloverDb means a subdirectory for DBdata. This subdirectory will be created in the directory which is set as derby.system.home (or in the workingdirectory if derby.system.home is not set). Value databases/cloverDb is a default value, you maychange it.

MySQL

CloverETL Server requires MySql 5.x


jdbc.driverClassName=com.mysql.jdbc.Driverjdbc.url=jdbc:mysql://127.0.0.1:3306/clover?useUnicode=true&characterEncoding=utf8jdbc.username=rootjdbc.password=jdbc.dialect=org.hibernate.dialect.MySQLDialect

Since 3.0 JDBC driver is not included in CloverETL Server web archive, thus it must be added to the applicationserver classpath.

Create DB with proper charset, like this:

CREATE DATABASE IF NOT EXISTS clover DEFAULT CHARACTER SET 'utf8';

DB2

DB2 on Linux/Windows


jdbc.driverClassName=com.ibm.db2.jcc.DB2Driverjdbc.url= jdbc:db2://localhost:50000/cloverjdbc.username=usrjdbc.password=pwdjdbc.dialect=org.hibernate.dialect.DB2Dialect


99

Possible problems

Wrong pagesize

Database clover has to be created with suitable PAGESIZE. DB2 has several possible values for this property:4096, 8192, 16384 or 32768.

CloverETL Server should work on DB with PAGESIZE set to 16384 or 32768. If PAGESIZE value is not setproperly, there should be error message in the log file after failed CloverETL Server startup:

ERROR:DB2 SQL Error: SQLCODE=-286, SQLSTATE=42727, SQLERRMC=16384; ROOT, DRIVER=3.50.152

SQLERRMC contains suitable value for PAGESIZE.

You can create database with proper PAGESIZE like this:

CREATE DB clover PAGESIZE 32768;

The table is in the reorg pending state

After some ALTER TABLE commands, some tables may be in "reorg pending state". This behaviour is specific forDB2. ALTER TABLE DDL commands are executed only during the first start of new CloverETL Server version.

Error message for this issue may look like this:

Operation not allowed for reason code "7" on table "DB2INST2.RUN_RECORD".. SQLCODE=-668, SQLSTATE=57016

or like this

DB2 SQL Error: SQLCODE=-668, SQLSTATE=57016, SQLERRMC=7;DB2INST2.RUN_RECORD, DRIVER=3.50.152

In this case "RUN_RECORD" is table name which is in "reorg pending state" and "DB2INST2" is DB instancename.

To solve this, go to DB2 console and execute command (for table run_record):

reorg table run_record

DB2 console output should look like this:

db2 => connect to clover1Database Connection Information Database server = DB2/LINUX 9.7.0SQL authorization ID = DB2INST2Local database alias = CLOVER1 db2 => reorg table run_recordDB20000I The REORG command completed successfully.db2 => disconnect clover1DB20000I The SQL DISCONNECT command completed successfully.

"clover1" is DB name

DB2 does not allow ALTER TABLE which trims DB column length.

This problem depends on DB2 configuration and we've experienced this only on some AS400s so far. CloverETLServer applies set of DP patches during the first installation after application upgrade. Some of these patches


100

may apply column modifications which trims length of the text columns. These changes never truncate any data,however DB2 does not allow this since it "may" truncate some data. DB2 refuses these changes even in DB tablewhich is empty. Solution is, to disable the DB2 warning for data truncation, restart CloverETL Server whichapplies patches, then enable DB2 warning again.

DB2 on AS/400

The connection on AS/400 might be slightly different.


jdbc.driverClassName=com.ibm.as400.access.AS400JDBCDriverjdbc.username=javlinjdbc.password=cloverjdbc.url=jdbc:as400://host/cloversrv;libraries=cloversrv;date format=isojdbc.dialect=org.hibernate.dialect.DB2400Dialect

Use credentials of your OS user for jdbc.username and jdbc.password.

cloversrv in jdbc.url above is the name of the DB schema.

You can create schema in AS/400 console:

• execute command STRSQL (SQL console)

• execute CREATE COLLECTION cloversrv IN ASP 1

• cloversrv is the name of the DB schema and it may be at most 10 characters long

Proper JDBC driver must be in the application server classpath.

I use JDBC driver jt400ntv.jar, which I've found in /QIBM/ProdData/Java400 on the server.

Use jt400ntv.jar JDBC driver.

Do not forget to add jar with JDBC driver to the Tomcat classpath.

Oracle


jdbc.driverClassName=oracle.jdbc.OracleDriverjdbc.url=jdbc:oracle:thin:@host:1521:dbjdbc.username=userjdbc.password=passjdbc.dialect=org.hibernate.dialect.Oracle9Dialect

Do not forget to add jar with JDBC driver to the application server classpath.

Since CloverETL Server version 1.2.1, dialect org.hibernate.dialect.Oracle10gDialect is nolonger available. Please use org.hibernate.dialect.Oracle9Dialect instead.

These are privileges which have to be granted to schema used by CloverETL Server:

CONNECTCREATE SESSIONCREATE/ALTER/DROP TABLECREATE/ALTER/DROP SEQUENCE


101

QUOTA UNLIMITED ON <user_tablespace>;QUOTA UNLIMITED ON <temp_tablespace>;

MS SQL

Ms SQL requires configuration of DB server.

• Allowing of TCP/IP connection:

• execute tool SQL Server Configuration Manager

• go to Client protocols

• switch on TCP/IP (default port is 1433)

• execute tool SQL Server Management Studio

• go to Databases and create DB clover

• go to Security/Logins and create user and assign this user as owner of DB clover

• go to Security and check SQL server and Windows authentication mode


jdbc.driverClassName=com.microsoft.sqlserver.jdbc.SQLServerDriverjdbc.url=jdbc:sqlserver://localhost:1433;databaseName=cloverjdbc.username=userjdbc.password=passjdbc.dialect=org.hibernate.dialect.SybaseDialect


Postgre SQL


jdbc.driverClassName=com.microsoft.sqlserver.jdbc.SQLServerDriverjdbc.url=jdbc:postgresql://localhost/clover?charSet=UTF-8jdbc.username=postgresjdbc.password=jdbc.dialect=org.hibernate.dialect.PostgreSQLDialect


JNDI DB DataSource

Server can connect to JNDI DB DataSource, which is configured in application server or container. However thereare some CloverETL parameters which must be set, otherwise the behaviour may be unpredictable:

datasource.type=JNDI # type of datasource; must be set, because default value is JDBCdatasource.jndiName=# JNDI location of DB DataSource; default value is java:comp/env/jdbc/clover_server #jdbc.dialect=# Set dialect according to DB which DataSource is connected to. The same dialect as in sections above. #

The parameters above may be set in the same ways as other params (in properties file or Tomcat context file)


102

Example of DataSource configuration in Apache Tomcat. Add following code to context file.

<Resource name="jdbc/clover_server" auth="Container" type="javax.sql.DataSource" driverClassName="com.mysql.jdbc.Driver" url="jdbc:mysql://192.168.1.100:3306/clover?useUnicode=true&characterEncoding=utf8" username="root" password="" maxActive="20" maxIdle="10" maxWait="-1"/>

Note

Special characters you type in the context file have to be specified as XML entities. E.g. ampersand"&" as "&" etc.

List of Properties

Table 18.1. General configuration

key description default

config.file location of CloverETL Server configuration file [working_dir]/cloverServer.properties

license.file location of CloverETL Server licence file (license.dat)

engine.config.file location of CloverETL engine configuration propertiesfile

properties file packedwith CloverETL

private.properties List of server properties which are used only byCloverETL Server code. So these properties are notaccessible outside of the ServerFacade. By default thereare all properties which may contain password in the list.Basically it means, that their values are not visible forweb GUI users. Values are replaced by single star "*".Changes in this list may cause unexpected behavior ofsome server API.

jdbc.password,executor.password,security.ldap.password,clover.smtp.password

engine.plugins.src This property may contain absolute path to some"source" of additional CloverETL engine plugins. Theseplugins are not a substitute for plugins packed in WAR."Source" may be directory or zip file. Both directory andzip must contain subdirectory for each plugin. Changesin the directory or the ZIP file apply only when the serveris restarted. For details see section "Extensibility - engineplugins".

empty

datasource.type Set this explicitly to JNDI if you need CloverETL Serverto connect to DB using JNDI datasource. In such case,parameters "datasource.jndiName" and "jdbc.dialect"must be set properly. Possible values: JNDI | JDBC

JDBC

datasource.jndiName JNDI location of DB DataSource. It is applied only if"datasource.type" is set to "JNDI".

java:comp/env/jdbc/clover_server

jdbc.driverClassName class name for jdbc driver name

jdbc.url jdbc url used by CloverETL Server to store data

jdbc.username jdbc database user name

jdbc.password jdbc database user name

jdbc.dialect hibernate dialect to use in ORM

quartz.driverDelegateClass SQL dialect for quartz. Value is automatically derivedfrom "jdbc.dialect" property value.


103


security_enabled true | false If it is set to false, then no authentication isrequired and anyone has admin privileges.

true

security.default_domain Domain which all new users are included in. Stored inuser's record in the database. Shouldn't be changed unlessthe "clover" must be white-labelled.

clover

security.basic_authentication.features_list

Semi-colon separated list of features which are accessibleusing HTTP and which should be protected by BasicHTTP Authentication. Each feature is specified by itsservlet path.

/request_processor;/simpleHttpApi;/launch;/launchIt;/downloadStorage;/downloadFile;/uploadSandboxFile;/downloadLog

security.basic_authentication.realm

Realm string for HTTP Basic Authentication. CloverETL Server

security.digest_authentication.features_list

Semi-colon separated list of features which are accessibleusing HTTP and which should be protected by HTTPDigest Authentication. Each feature is specified by itsservlet path. Please keep in mind, that HTTP DigestAuthentication is feature added to the version 3.1. Ifyou upgraded your older CloverETL Server distribution,users created before the upgrade cannot use the HTTPDigest Authentication until they reset their passwords. Sowhen they reset their passwords (or the admin does it forthem), they can use Digest Authentication as well as newusers.

/webdav

security.digest_authentication.realm

Realm string for HTTP Digest Authentication. If itis changed, all users have to reset their passwords,otherwise they won't be able to access to the serverfeatures protected by HTTP digest Authentication.

CloverETL Server

security.digest_authentication.nonce_validity

Interval of validity for HTTP Digest Authenticationspecified in seconds. When the interval passes, serverrequires new authentication from the client. Most of theHTTP clients do it automatically.

300

clover.event.fileCheckMinInterval

Interval of file checkes (in milliseconds) See Chapter 12,File event listeners (p. 71) for details.

1000

clover.smtp.host SMTP server hostname or IP address

clover.smtp.port SMTP server port

clover.smtp.authentication true/false If it is false, username and password areignored

clover.smtp.username SMTP server username

clover.smtp.password SMTP server password

logging.project_name used in log messages where it is neccessary to name theproduct name

CloverETL

logging.default_subdir name of default subdirectory for all server logs; itis relative to the path specified by system property"java.io.tmpdir". Don't specify absolute path, useproperties which are intended for absolute path.

cloverlogs


104


launch.log.dir Location, where server should store launch requests logs.See Launch Services section for details.

${java.io.tmpdir}/[logging.default_subdir]/launch where${java.io.tmpdir} issystem property

graph.logs_path Location, where server should store Graph run logs. SeeLogging section for details.

${java.io.tmpdir}/[logging.default_subdir]/graphwhere${java.io.tmpdir} issystem property

temp.default_subdir Name of default subdirectory for server tmp files; itis relative to the path specified by system property"java.io.tmpdir".

clovertmp

graph.debug_path Location, where server should store Graph debug info. ${java.io.tmpdir}/[temp.default_subdir]/debug where${java.io.tmpdir} issystem property

graph.dictionary_path Location, where server should store graph dictionarytemporary files.

${java.io.tmpdir}/[temp.default_subdir]/dictionary where${java.io.tmpdir} issystem property

graph.pass_event_params_to_graph_in_old_style

Since 3.0. It is switch for backwards compatibility ofpassing parameters to the graph executed by graph event.In version prior to 3.0 all params has been passed toexecuted graph. Since 3.0 just specified parameters arepassed. Please see Task - Execution of Graph (p. 48) fordetails.

false

threadManager.pool.corePoolSize

Number of threads which are always active (runningor idling). Related to thread pool for processing serverevents.

4

threadManager.pool.queueCapacity

Max size of the queue(FIFO) which contains taskswaiting for thread. Related to thread pool for processingserver events. It means, that there won't be more then"queueCapacity" waiting tasks. i.e. queueCapacity=0 -no waiting tasks, each task is immediatelly executed inavailable thread or in new thread. queueCapacity=1024 -up to 1024 tasks may be waiting in the queue for availablethread from "corePoolSize".

12

threadManager.pool.maxPoolSize

Max number of active threads. If no thread from corepool is available and queue capacity is exceeded, poolcreates new threads up to "maxPoolSize" threads. Ifthere are more concurrent tasks then maxPoolSize, threadmanager refuses to execute it. Thus keep queueCapacityor maxPoolSize big enough.

1024

task.archivator.batch_size Max number of records deleted in one batch. It is usedfor deleting of archived run records.

50

launch.http_header_prefix Prefix of HTTP headers added by launch services to theHTTP response.

X-cloveretl


105


task.archivator.archive_file_prefix

Prefix of archive files created by archivator. cloverArchive_

license.context_names Comma separated list of web-app contexts which maycontain license. Each of them has to start with slash!Works only on Apache Tomcat.

/clover-license,/clover_license

license.display_header Switch which specifies whether display license header inserver web GUI or not.

false

Table 18.2. Defaults for job execution configuration - see section Job config properties for details


executor.tracking_interval Interval in milliseconds for scanning current status of runninggraph. The shorter interval, the bigger log file.

2000

executor.log_level Log level of graph runs. TRACE | DEBUG | INFO | WARN |ERROR

INFO

executor.max_running_concurrently Amount of graph instances which may exist(or run) concurrently.0 means no limits

0

executor.max_graph_instance_age Interval in milliseconds. Specifies how long graph instance canbe idling before it is released from memory. 0 means no limits.This property has been renamed since 2.8. Original name wasexecutor.maxGraphInstanceAge

0

executor.classpath Classpath for transformation/processor classes used in the graph.Directory [sandbox_root]/trans/ does not have to be listed here,since it is automatically added to graph run classpath.

executor.skip_check_config Disables check of graph configuration. Increases performanceof graph execution, however may be useful during graphdevelopment.

true

executor.password Password for decoding of encoded DB connection passwords.

executor.verbose_mode If true, more descriptive logs of graph runs are generated. true

executor.use_jmx If true, graph executor registers jmx mBean of running graph. true

executor.debug_mode If true, edges with enabled debug store data into files in debugdirectory. See property "graph.debug_path"

false

See "Clustering" section for more properties.

106

Chapter 19. Graph parametersCloverETL Server passes set of parameters for each graph execution. Please keep in mind, that placeholders${paramName} are resolved only during loading of graph XML, so if you need placeholders resolving for eachgraph execution, graph cannot be pooled. However current parameter values are always accessible by inline javacode like this:

String runId = getGraph().getGraphProperties().getProperty("RUN_ID");

Properties may be added or replaced like this:

getGraph().getGraphProperties().setProperty("new_property", value );

This is set of parameters which are always set by CloverETL Server:

Table 19.1. Defaults for graph execution configuration - see section Graph config properties for details

key description

SANDBOX_CODE Code of sandbox which contains executed graph.

JOB_FILE Path to the file, relative to sandbox root path.

SANDBOX_ROOT Absolute path sandbox root.

RUN_ID ID of the graph execution. In standalone mode or incluster mode, it is always unique. It may be lower then0 value, if the run record isn't persistent. See "LaunchServices" for details.

Another sets of parameters according the type of execution

There are some more parameters in dependence of way, how the graph is executed.

executed from Web GUI

no more parameters

executed by Launch Service invocation

Service parameters which have attribute Pass to graph enabled are passed to the graph not only as "dictionary"input data, but also as graph parameter.

executed by HTTP API run graph operation invocation

Any URL parameter with "param_" prefix is passed to executed graph but without "param_" prefix. i.e."param_FILE_NAME" specified in URL is paased to the graph as property named "FILE_NAME".

executed by RunGraph component

Since 3.0 only parameters specified by "paramsToPass" attribute are passed from the "parent" graph to the executedgraph. However common properties (RUN_ID, PROJECT_DIR, etc.) are overwritten by new values.

Chapter 19. Graph parameters

107

executed by WS API method executeGraph invocation

Parameters with values may be passed to the graph with the request for execution.

executed by task "graph execution" by scheduler

Table 19.2. passed parameters

key description

EVENT_SCHEDULE_EVENT_TYPE Type of schedule SCHEDULE_PERIODIC |SCHEDULE_ONETIME

EVENT_SCHEDULE_LAST_EVENT Date/time of previous event

EVENT_SCHEDULE_DESCRIPTION Schedule description, which is displayed in web GUI

EVENT_USERNAME User who "owns" the event. For schedule it is the userwho created the schedule

EVENT_SCHEDULE_ID ID of schedule which triggered the graph

executed from JMS listener

There are many graph parameters and dictionary entries passed, depending on the type of incomming message.See details in Chapter 9, JMS messages listeners (p. 65).

executed by task "graph execution" by graph event listener

Since 3.0 only specified properties from "source" graph are passed to executed graph by default. There is serverconfig property "graph.pass_event_params_to_graph_in_old_style" which can change this behavior so that ALLparameters from "source" graph are passed to the executed graph. This switch is implemented for backwardscompatibility. Regarding the default behaviour: You can specified list of parameters to pass in the editor of graphevent listener. Please see the section "Task - Execution of Graph" for details.

However following parameters with current values are always passed to the target graph


key description

EVENT_RUN_SANDBOX Sandbox with graph, which is source of the event

EVENT_GRAPH_EVENT_TYPE GRAPH_STARTED | GRAPH_FINISHED |GRAPH_ERROR | GRAPH_ABORTED |GRAPH_TIMEOUT |GRAPH_STATUS_UNKNOWN

EVENT_RUN_GRAPH graphId of the graph, which is source of the event

EVENT_RUN_ID ID of the graph execution, which is source of the event.

EVENT_TIMEOUT Number of miliseconds which specifies interval oftimeout. Makes sence only for "timeout" graph event.

EVENT_RUN_RESULT Result (or current status) of the execution, which issource of the event.

EVENT_USERNAME User who "owns" the event. For graph events it is theuser who created the graph event listener

Chapter 19. Graph parameters

108

executed by task "graph execution" by file event listener


key description

EVENT_FILE_PATH Path to file, which is source of the event. Does notcontain file name. Does not end with file separator.

EVENT_FILE_NAME Filename of the file which is source of the event.

EVENT_FILE_EVENT_TYPE SIZE | CHANGE_TIME | APPEARANCE |DISAPPEARANCE

EVENT_FILE_PATTERN Pattern specified in file event listener

EVENT_FILE_LISTENER_ID

EVENT_USERNAME User who "owns" the event. For file events it is the userwho created the file event listener

How to add another graph parameters

Additional "Graph Config Parameters"

It is possible to add so-called additional parameters in Web GUI - section Sandboxes for the selected graph or forall graphs in the selected sandbox. See details in the section called “Job config properties” (p. 32).

Task "execute_graph" parameters

The "execute graph" task may be triggered by schedule, graph event listener or file event listener. Task editorallows you to specify key=value pairs which are passed to executed graph.

109

Chapter 20. Recommendations for transformationsdevelopers

Add external libraries to app-server classpath

i.e. connections (JDBC/JMS) may require third party libraries. It is strongly recommended to add these librariesto app-server classpath.

CloverETL allows you to specify these libraries directly in graph definition so CloverETL may load theselibraries dynamically, but external libraries may cause memory leak resulting with "java.lang.OutOfMemoryError:PermGen space" in this case.

In addition, app-servers should have the JMS API on their classpath and the third-party libraries often bundle thisAPI as well. So it may result in classloading conflicts if these libraries are not loaded by the same classloader.

Another graphs executed by RunGraph component may beexecuted only in the same JVM instance

In server environment, all graphs are executed in the same VM instance. Attribute "same instance" of RunGraphcomponent cannot be set to false.

110

Chapter 21. Logging

Main logs

CloverETL Server uses log4j library for logging. WAR file contains default log4j configuration.

By default, log files are produced in directory specified by system property "java.io.tmpdir" in "cloverlogs"subdirectory.

"java.io.tmpdir" usually contains common system temp dir i.e. "/tmp". On tomcat, it is usually"[TOMCAT_HOME]/temp"

Default logging configuration may be overridden by system property "log4j.configuration", which should containURL to log4j config file.

log4j.configuration=file:/home/clover/config/log4j.xml

Since such configuration overrides default configuration, it may have influence over Graph run logs. So your ownlog config has to contain following fragment to preserve Graph run logs

<logger name="Tracking" additivity="false"> <level value="debug"/></logger>

These system properties allow logging of HTTP requests/responses to stdout:

client side:

com.sun.xml.ws.transport.http.client.HttpTransportPipe.dump=true (for moreinformation consult CloverETL Designer Users's Guide - chapter Integrating CloverETL Designer withCloverETL Server)

server side:

com.sun.xml.ws.transport.http.HttpAdapter.dump=tru

Graph run logs

Each graph run has it is own log file, which is accessible i.e. in web GUI, section "executions history".

By default these log files are produced in subdirectory cloverLogs/graph in the directory specified by"java.io.tmpdir" system property.

It is possible to specify different location for these logs by CloverETL property "graph.logs_path". This propertydoes not have any influence over main server logs.

http://www.cloveretl.com/documentation/UserGuide/topic/com.cloveretl.gui.docs/docs/designer-server-integration.html

http://www.cloveretl.com/documentation/UserGuide/topic/com.cloveretl.gui.docs/docs/designer-server-integration.html

111

Chapter 22. Extensibility (Embedded OSGiframework)CloverETL Server implements extensibility of its APIs, so the server may expose its fetures with custom API.

For now, there are two possibilities: Groovy code API and OSGi plugin.

Groovy Code API

Since 3.3

CloverETL Server Groovy Code API allows clients to execute groovy code stored on the server by HTTP request.Executed code has access to the serverFacade, instance HTTP request and HTTP response, so it's possible toimplement custom CloverServer API in the Groovy code.

To execute the code call URL:

http://[host]:[port]/clover/groovy/[sandboxCode]/[pathToGroovyCodeFile]

Protocol HTTP can be changed to secured HTTPS according to web server configuration.

Server uses Basic or Digest authentication according to the configuration. so the user must be authorized and musthave permission to execute in the specified sandbox and permission to call "Groovy Code API".

Please note, that permission to call "Groovy Code API" (and edit them) is wery strong permission, since the GroovyCode can basically do the same as Java code and it's running as the same system process as whole applicationcontainer.

Variables accessible in the Groovy code

By default, there are some variables accessible in the groovy code



javax.servlet.http.HttpServletRequest request Instance of HTTP request, which triggered thecode.

javax.servlet.http.HttpServletResponse response Instance of HTTP response, which will be sentto the client when the script finishes.

javax.servlet.http.HttpSession session Instance of HTTP session.

javax.servlet.ServletConfig servletConfig instance of ServletConfig


com.cloveretl.server.api.ServerFacade serverFacade Instance of serverFacade usable for callingCloverETL Server core features.

WAR file contains JavaDoc of facade API andit is accessible on URL: http://host:port/clover/javadoc/index.html




Chapter 22. Extensibility(Embedded OSGi framework)

112

Code examples

Code may return string which will be returned as a content of HTTP response, or it may itself construct the outputto the output Writer

Following script writes its own output and doesn't return anything, so the underlaying servlet doesn't modifythe output at all. Advantage of this aproach is, that output may be constructed on the fly and sent to the clientcontinuously. However when the output stream(or writer) is opened, servlet won't send any error description incase of any error during the script processing.

response.getWriter().write("write anything to the output");

Following script returns String, so the underlaying servlet puts the string to the output. Advantage od this aproachis, that in case of any error during code processing, servlet returns full stacktrace, so the script may be fixed.However the constructed output may consume some amount of memory.

String output = "write anything to the output";return output;

Following script is little bit more complex. It returns XML with list of all configured schedules. User must havepermission to list the schedules.

// uses variables: response, sessionToken, serverFacadeimport java.io.*;import java.util.*;import javax.xml.bind.*;import com.cloveretl.server.facade.api.*;import com.cloveretl.server.persistent.*;import com.cloveretl.server.persistent.jaxb.*;

JAXBContext jc = JAXBContext.newInstance( "com.cloveretl.server.persistent:com.cloveretl.server.persistent.jaxb" );Marshaller m = jc.createMarshaller();m.setProperty(Marshaller.JAXB_ENCODING, "UTF-8");m.setProperty(Marshaller.JAXB_FORMATTED_OUTPUT, Boolean.TRUE);m.setProperty(Marshaller.JAXB_SCHEMA_LOCATION, "/clover/schemas/csc.xsd");

Response<List<Schedule>> list = serverFacade.findScheduleList(sessionToken, null);SchedulesList xmlList = new SchedulesList();xmlList.setSchedulesList(list.getBean());m.marshal(xmlList, response.getWriter());

Embedded OSGi framework

Since 3.0

CloverETL Server includes embedded OSGi framework which allows implementation of "plugin" (OSGi bundle)which works as new API (or even GUI) of the server and it is independent of released clover.war.

Plugin possibilities

Basically the plugin may work as new server API similarly as Launch Services, HTTP API, WebServices API.It may be just simple JSP, HttpServlet or complex SOAP Web Services. So if the plugin contains some HTTPservice, it is registered to listen on specified URL during the startup and incoming HTTP requests are "bridged"

Chapter 22. Extensibility(Embedded OSGi framework)

113

from the web container to the plugin. Plugin itself has access to the internal CloverETL Server interface called"ServerFacade". ServerFacade offers methods for execution graphs, obtaining of graph status and executionshistory, manipulation with scheduling, listeners, configuration and many more. So the API may be customizedaccording to the needs of specific deployment.

Deploying an OSGi bundle

There are 2 CloverETL Server configuration properties related to the OSGi framework.

• plugins.path - Absolute path to the directory containing all your plugins (jar files).

• plugins.autostart - It is comma separated plugin names list. These plugins will be started during server startup.Theoretically OSGi framework can start the OSGi bundle on demand, however it is unreliable when the servletbridge to the servlet container is used, so it is strongly recommended to name all your plugins.

So do deploy your plugin: set two config properties, copy plugin to the directory specified by "plugins.path" andrestart the server.

114

Chapter 23. Extensibility CloverETL engine pluginsSince 3.1.2

CloverETL Server may use external engine plugins loaded from specified source. Source is specified by configproperty "engine.plugins.src"

See details about possibilities of CloverETL configuration in Chapter 18, Configuration (p. 96)

This property must be absolute path to the directory or zip file with additional CloverETL engine plugins. Bothdirectory and zip must contain subdirectory for each plugin. These plugins are not a substitute for plugins packedin WAR. Changes in the directory or the ZIP file apply only when the server is restarted.

Each plugin has its own class-loader which uses parent-first strategy by default. Parent of plugins' classloadersis web-app classloader (content of [WAR]/WEB-INF/lib). If the plugin uses any third-party libraries, there maybe some conflict with libraries on parent-classloaders classpath. These are common exceptions/errors suggesting,that there is something wrong with classloading:

• java.lang.ClassCastException

• java.lang.ClassNotFoundException

• java.lang.NoClassDefFoundError

• java.lang.LinkageError

There are couple of ways how to ged rid of such conflicts:

• Remove your conflicting third-party libraries and use libraries on parent classloaders (web-app or app-serverclassloaders)

• Use different class-loading strategy for your plugin.

• in the plugin descriptor plugin.xml, set attribute greedyClassLoader="true" in the element "plugin"

• it means, that plugin classloader will use self-first strategy

• Set inverse class-loading strategy for selected java packages.

• In the plugin descriptor plugin.xml, set attribute "excludedPackages" in the element "plugin".

• It's comma separated list of package prefixes. E.g. like this:excludedPackages="some.java.package,some.another.package"

• In previous example all classes from "some.java.package", "some.another.package" and all their sub-packages would be loaded with the inverse loading strategy then the rest of classes on the plugins classpath.

Of course, the suggestions above may be combined. It's not easy to find the best solution for these conflicts andit may depend on the libraries on app-server classpath.

For more convinient debugging it is useful to set TRACE log level for related class-loaders.

<logger name="org.jetel.util.classloader.GreedyURLClassLoader"> <level value="trace"/></logger><logger name="org.jetel.plugin.PluginClassLoader"> <level value="trace"/></logger>

See "Logging" section for details about overriding server log4j configuration.

115

Chapter 24. ClusteringCloverETL Server only works in the cluster if the user's license allows it.

There are two common cluster features, high availability and scalability. Both of them are implemented byCloverETL Server on various levels. This section should clarify the basics of CloverETL Clustering.

High Availability

Since version 3.0, CloverETL Server does not recognize any differences between cluster nodes. Thus, there areno "master" or "slave" nodes meaning all nodes can be virtually equal. There is no single point of failure(SPOF)in the CloverETL cluster itself, however SPOFs may be in the input data or some other external element.

Clustering offers high availability(HA) for all features accessible through HTTP. This includes sandbox browsing,modification of services configuration (scheduling, launch services, listeners) and primarily graph executions.Any cluster node may accept incoming HTTP requests and process them itself or delegate it to another node.

Since all nodes are equal, almost all requests may be processed by any cluster node:

• All graph files, metadata files, etc. are located in shared sandboxes. Thus all nodes have access to them. Ashared filesystem may be a SPOF, thus it is recommended to use a replicated filesystem instead.

• The database is shared by all cluster nodes. Again, a shared DB might be a SPOF, however it may be clusteredas well.

But there is still a possibility, that a node cannot process a request by itself. In such cases, it completely andtransparently delegates the request to a node which can process the request.

These are the requests which are limited to one (or more) node(s):

• a request for the content of a partitioned or local sandbox. These sandboxes aren't shared among all clusternodes. Please note that this request may come to any cluster node which then delegates it to a target node,however, this target node must be up and running.

• A graph is configured to use a partitioned or local sandbox. These graphs need nodes which have a physicalaccess to the required sandboxes.

Thus an inaccessible partitioned or local sandbox may cause a failure from the request, however...

1. it is still possible to configure redundant sandboxes stored on other cluster nodes.

2. these types of sandboxes are used only for scalability on the data level(described below), which is a differentapproach to using a CloverETL cluster.

CloverETL itself implements a load balancer for executing graphs. So a graph which isn't configured for somespecific node(s) may be executed anywhere in the cluster and the CloverETL load balancer decides, according tothe current load, which node will process the graph. All this is done transparently.

To achieve HA, it is recommended to use an independent HTTP load balancer. Independent HTTP load balancersallow transparent fail-overs for HTTP requests. They send requests to the nodes which are running.

Scalability

There are two independent levels of scalability implemented. Scalability of transformation requests(and any HTTPrequests) and data scalability (parallel data processing).

Both of these "scalability levels" are "horizontal". Horizontal scalability means adding nodes to the cluster, whereasvertical scalability means adding resources to a single node. Vertical scalability is supported natively by theCloverETL engine and it is not described here.

Chapter 24. Clustering

116

Transformation Requests

Basically, the more nodes we have in the cluster , the more transformation requests (or HTTP requests in general)we can process at one time. This type of scalability is the CloverETL server's ability to support a growing numberof clients. This feature is closely related to the use of an HTTP load balancer which is mentioned in the previoussection.

Parallel Data Processing

When a transformation is processed in parallel, the whole graph (or its parts) runs in parallel on multiple clusternodes having each node process just a part of the data.

So the more nodes we have in the cluster, the more data can be processed in the specified time.

The data may be split(partitioned) before the graph execution or by the graph itself on the fly. The resulting datamay be stored in partitions or gathered and stored as one group of data.

The curve of scalability may differ according to the type of transformation. It may be almost linear, which isalmost always ideal, except when there is a single data source which cannot be read by multiple readers in parallellimiting the speed of further data transformation. In such cases it is not beneficial to have parallel data processingsince it would actually wait for input data.

Node Allocation

Node allocation is the specification of which cluster nodes will run the graph and which parts of the graph theywill run. Allocation is basically specified by the partitioned sandboxes used in the graph phase. Each phase mayhave its own (just one) allocation. Basically, each partitioned sandbox has a list of locations. When some part ofthe graph runs in parallel, there is one worker for each partitioned sandbox location. See "Partitioned sandbox" inPartitioned and Local Sandboxes (p. 118) for details.

Allocation is specified in the graph either by:

• sandbox resources pointing to a partitioned sandbox, if workers read/write some partitioned data to/from theirown location of this partitioned sandbox, or by

• the node attribute "node allocation", if workers do not read/write their partitioned data, however there must bean allocation specified.

If there is a conflict, execution fails and an error message appears containing the description of the conflict. Asingle conflict may be caused by using two different allocations in a single phase.

Partitioning/gathering Data

As mentioned before, data may be partitioned and gathered in multiple ways. It may be prepared before the graphis executed or it may be partitioned on the fly.

Partitioning/gathering "on the fly"

There are two special components to consider: ClusterPartitioner and ClusterGather. Both work similarly, but inthe opposite way.

ClusterPartitioner works like a common partitioner, but node allocation is applied simultaneously behind theClusterPartitioner component. All components preceding the ClusterPartitioner run on just one node (so calledthe primary worker - see below) whereas components behind the ClusterPartitioner run in parallel according tonode allocation. Thus, these nodes work with just part of the data. There are more partitioning types: "round-robin" (default), "by record key", and "by load".


117

ClusterGather works in the opposite way. Components preceding the gather run in parallel while componentsbehind the gather run on just one node (primary worker). The cluster gather component gathers records in thesame way as SimpleGather and its attributes are the same. By default it does not sort input records in any way.It just gathers them in the order they come.

Primary worker node - some parts of the graph designed to run in parallel may run on a single node anyway. i.e. thepart where the graph reads/writes data from/to a single resource. It may be the part preceding ClusterPartitioneror the part behind ClusterGatherer respectively. It also may be on all components in the phase which do nothave node allocation specified at all. Each phase may have its own primary worker. All graph primary workersare chosen during graph execution primarily according to the local sandbox datasources used in the phases.Basically, the node which has direct(local) access to a sandbox datasource(s) used in the phase is selected as theprimary worker. Of course, there may be multiple different local sandbox datasources, or even no local sandboxdatasources used in the phase. In such cases, the server uses some minor parameters to choose the primary worker.

Both components may be combined in a single phase in any way, but there must be just one node allocation andjust one primary worker in each single phase.

This example shows how data would be processed in 2 different node allocations, on 2 different primary workers.

• phase 1 starts

• processing data on primary worker (nodeA)

• cluster partitioner component

• processing data in parallel (nodeA, nodeB, nodeC)

• cluster gatherer component


• phase 1 ends

• phase 2 starts


• cluster partitioner component

• processing data in parallel (nodeB, nodeD)

• phase 2 ends

• phase 3 starts

• processing data in parallel (nodeB, nodeD)


• processing data on primary worker (nodeD)

• phase 3 ends

Results are stored on a different node (nodeD) then the node that read (nodeA) and data is actually repartitioned(from nodeA, nodeB, nodeC to nodeB, nodeD).

Partitioning/gathering data by external tools

Partitioning data on the fly may in some cases be an unnecessary bottleneck. Splitting data using low-level toolscan be much better for scalability. The optimal case being, that each running worker reads data from an independent


118

data source. Thus there does not have to be a ClusterPartitioner component in the first phase and the graph runsin parallel from the beginning.

• phase 1 starts




• phase 1 ends

Or the whole graph may run in parallel, however the results would be partitioned.

• phase 1 starts


• phase 1 ends

Partitioned and Local Sandboxes

Partitioned and local sandboxes were mentioned in previous sections. These new sandbox types were introducedin version 3.0 and they are vital for parallel data processing.

Together with shared sandboxes, we have three sandbox types in total.

Shared sandbox

This type of sandbox must be used for all data which is supposed to be accessible on all cluster nodes. This includesall graphs, metadata, connections, classes and input/output data for graphs which should support HA, as describedabove.

Figure 24.1. Dialog form for creating new shared sandbox

As you can see in the screenshot above, you cannot specify any root path on the filesystem. Shared sandboxesare stored in the directory specified by "cluster.shared_sandboxes_path". Each shared sandbox has its ownsubdirectory in it, which is named by sandbox ID.

Local sandbox

This sandbox type is intended for data, which is accessible only by certain cluster nodes. It may include massiveinput/output files. The purpose being, that any cluster node may access content of this type of sandbox, but onlyone has local(fast) access and this node must be up and running to provide data. The graph may use resources frommultiple sandboxes which are physically stored on different nodes since cluster nodes are able to create networkstreams transparently as if the resource was a local file. See Using a Sandbox Resource as a Component DataSource (p. 119) for details.


119

Do not use local sandbox for common project data (graphs, metadata, connections, lookups, properties files, etc.).It would cause odd behaviour. Use shared sandboxes instead.

Figure 24.2. Dialog form for creating new local sandbox

Partitioned sandbox

This type of sandbox is actually an abstract wrapper for a couple of physical locations existing typically on differentcluster nodes. However, there may be multiple locations on the same node. A partitioned sandbox has two purposeswhich are both closely related to parallel data processing.

1. node allocation specification - locations of a partitioned sandbox define the workers which will run the graphor its parts. So each physical location will cause a single worker to run. This worker does not have to actuallystore any data to "its" location. It is just a way to tell the CloverETL Server: "execute this graph/phase in parallelon these nodes"

2. storage for part of the data during parallel data processing. Each physical location contains only part of thedata. In a typical use, we have input data split in more input files, so we put each file into a different locationand each worker processes its own file.

As you can see on the screenshot above, for a partitioned sandbox, you can specify one or more physical locationson different cluster nodes.

Do not use partitioned sandbox for common project data (graphs, metadata, connections, lookups, properties files,etc.). It would cause odd behavior. Use shared sandboxes instead.

Using a Sandbox Resource as a Component Data Source

A sandbox resource, whether it is a shared, local or partitioned sandbox (or ordinary sandbox on standalone server),is specified in the graph under the fileURL attributes as a so called sandbox URL like this:

sandbox://data/path/to/file/file.dat

where "data" is a code for sandbox and "path/to/file/file.dat" is the path to the resource from the sandbox root.URL is evaluated by CloverETL Server during graph execution and a component (reader or writer) obtains theopened stream from the server. This may be a stream to a local file or to some other remote resource. Thus, a graphdoes not have to run on the node which has local access to the resource. There may be more sandbox resourcesused in the graph and each of them may be on a different node. In such cases, CloverETL Server would choosethe node with the most local resources to minimalize remote streams.

The sandbox URL has a specific use for parallel data processing. When the sandbox URL with the resourcein a partitioned sandbox is used, that part of the graph/phase runs in parallel, according to the node allocationspecified by the list of partitioned sandbox locations. Thus, each worker has it is own local sandbox resource.CloverETL Server evaluates the sandbox URL on each worker and provides an open stream to a local resourceto the component.

The sandbox URL may be used on standalone server as well. It is excellent choice when graph referencessome resources from different sandboxes. It may be metadata, lookup definition or input/output data. Of course,referenced sandbox must be accessible for the user who executes the graph.


120

Recommendations for Cluster Deployment

1. All nodes in the cluster should have a synchronized system date-time.

2. All nodes share sandboxes stored on a shared or replicated filesystem. The filesystem shared among all nodesis single point of failure. Thus, the use of a replicated filesystem is strongly recommended.

3. All nodes share a DB, thus it must support transactions. I.e. The MySQL table engine, MyISAM, may causestrange behaviour because it is not transactional.

4. All nodes share a DB, which is a single point of failure. Use of a clustered DB is strongly recommended.

5. Configure the license as "license.file" for this property on Tomcat. Do not use clover_license.war.Tomcat loads web-apps in an unpredictable order and for the cluster, the license must be loaded beforeCloverETL Server itself.

Figure 24.3. List of nodes joined to the cluster

Example of Distributed Execution

The following diagram shows a transformation graph used for parsing invoices generated by a few cell phonenetwork providers in Czech Republic.


121

The size of these input files may be up to a few gigabytes, so it is very beneficial to design the graph to workin the cluster environment.

Details of the Example Transformation Design

Please note there is only one phase and there are four cluster components in the graph (highlighted by red border).These components define a point of change "node allocation", so the part of the graph demarcated by thesecomponents is highlighted by the red rectangle. This part of the graph performs data processing in parallel. Thismeans that the components inside the dotted rectangle run on cluster nodes according to the "node allocation" ofthat part of the graph.

The rest of the graph runs just on one node called "primary worker".

Specification of "node allocation"

Since there is only one phase, the whole graph has just one primary worker and only one node allocation.

• node allocation is applied for groups of components running in parallel (demarcated by the four clustercomponents)

• the outer part of the graph run on a single node - primary worker.

The primary worker is specified by the sandbox code used in the URLs of input data. The following dialog showsthe File URL value: "sandbox://data/path-to-csv-file", where "data" is the ID of the server sandbox containing thespecified file. And it is the "data" local sandbox which defines the primary worker in the graph.

The part of the graph demarcated by the four cluster components may have specified its allocation by the fileURL attribute as well, but this part does not work with files at all, so there is no file URL. Thus, we will usethe "allocation" attribute. Since all components in this part must have the same allocation, it is sufficient to setit only for one component.

Again, "dataPartitioned" in the following dialog is the sandbox ID.


122

Let's investigate our sandboxes. This project requires 3 sandboxes: "data", "dataPartitioned" and"PhoneChargesDistributed".

• data

• contains input and output data

• local sandbox (yellow folder), so it has only one physical location

• accessible only on node "i-4cc9733b" in the specified path

• dataPartitioned

• partitioned sandbox (red folder), so it has a list of physical locations on different nodes

• does not contain any data and since the graph does not read or write to this sandbox, it is used only for thedefinition of "nodes allocation"

• on the following figure, allocation is configured for two cluster nodes

• PhoneChargesDistributed

• common sandbox containing the graph file, metadata, and connections

• shared sandbox (blue folder), so all cluster nodes have access to the same files


123

If the graph was executed with the sandbox configuration of the previous figure, the node allocation would be:

• components which run only on primary worker, will run only on the "i-4cc9733b" node according to the "data"sandbox location.

• components with allocation according to the "dataPartitioned" sandbox will run on nodes "i-4cc9733b" and"i-52d05425".

Scalability of the Example Transformation

The example transformation has been tested in the Amazon Cloud environment with the following conditions forall executions:

• the same master node

• the same input data: 1,2 GB of input data, 27 million records

• three executions for each "node allocation"

• "node allocation" changed between every 2 executions

• all nodes has been of "c1.medium" type

We tested "node allocation" from 1 single node, all the way up to 8 nodes.

The following figure shows the functional dependence of run-time on the number of nodes in the cluster:


124

Figure 24.4. Cluster Scalability

The following figure shows the dependency of "speedup factor" on the number of nodes in the cluster. The speedupfactor is the ratio of the average runtime with one cluster node and the average runtime with x cluster nodes. Thus:

speedupFactor = avgRuntime(1 node) / avgRuntime(x nodes)

We can see, that the results are favourable up to 4 nodes. Each additional node still improves cluster performance,however the effect of the improvement decreases. Nine or more nodes in the cluster may even have a negativeeffect because their benefit for performance may be lost in the overhead with the management of these nodes.

These results are specific for each transformation, there may be a transformation with much a better or possiblyworse function curve.

Figure 24.5. Speedup factor

Table of measured runtimes:

nodes runtime 1 [s] runtime 2 [s] runtime 3 [s] averageruntime [s]

speedup factor

1 861 861 861 861 1

2 467 465 466 466 1.85

3 317 319 314 316.67 2.72

4 236 233 233 234 3.68

5 208 204 204 205.33 4.19

6 181 182 182 181.67 4.74

7 168 168 168 168 5.13

8 172 159 162 164.33 5.24


125

Cluster configuration

Cluster can work properly only if each node is properly configured. Clustering must be enabled, nodeID must beunique on each node, all nodes must have access to shared DB and shared sandboxes, and all properties for inter-node cooperation must be set according to network environment.

Properties and possible configuration are the following:

• Mandatory properties (p. 125)

• Optional properties (p. 126)

• Example of 2 node cluster configuration (p. 126)

• Load balancing properties (p. 127)

Mandatory properties

Besides mandatory cluster properties, you need to set license.file, database connection and other necessaryproperties which are not specifically related to the cluster environment.

Table 24.1. Mandatory properties - these properties must be properly set on each node of the cluster

property type default

cluster.enabled boolean false

description: switch whether server is connected to the cluster or not

cluster.node.id String node01

description: each cluster node must have unique ID

cluster.shared_sandboxes_path String, path

description: Path, where all shared sandboxes are stored on this node. If clusteris enabled, all sandboxes are shared, thus "rootPath" attribute of thesandbox is ignored. Path to the root directory of the sandbox isconstructed like this: [shared_sandboxes_path]/[sandboxID]

cluster.jgroups.bind_address String, IP address 127.0.0.1

description: IP address of ethernet interface, which is used for communication withanother cluster nodes. Necessary for inter-node messaging.

cluster.jgroups.start_port int, port 7800

description: Port where jGroups server listens for inter-node messages.

cluster.jgroups.tcpping.initial_hosts String, in format:"IPaddress1[port1],IPaddress2[port2]"

127.0.0.1[7800]

description: List of IP addresses(with ports) where we expect runningand listening nodes. It is related to another nodes"bind_address" and "start_port" properties. I.e. like this:bind_address1[start_port1],bind_address2[start_port2],... It is notnecessary to list all nodes of the cluster, but at least one of listed host:portmust be running. Necessary for inter-node messaging.

cluster.http.url String, URL http://localhost:8080/clover

description: URL to the root of web application of configured node. Necessary forinter-node cooperation. This value will be sent to all other nodes in thecluster to let them know how to connect to this node.


126

Optional properties

Table 24.2. Optional properties - these properties aren't vital for cluster configuration - default valuesare sufficient

property type default description

cluster.node.sendinfo.interval int 5000 time interval in ms; eachnode sends info about itselfto another nodes; this intervalspecified how often the info issent

cluster.node.remove.interval int 15000 time interval in ms; if no nodeinfo comes in this interval, nodeis considered as lost and it isremoved from the cluster

cluster.max_allowed_time_shift_between_nodes int 2000 Max allowed time shift betweennodes. If time shift exceeds this,node will be selected as invalid.

cluster.group.name String cloverCluster Each cluster has its uniquegroup name. If you need 2clusters in the same networkenvironment, each of themwould have its own groupname.

cluster.max_allowed_time_shift_between_nodes int 2000 How many miliseconds ismaximum allowed time shiftbetween nodes in the cluster.All nodes must have systemtime synchronized. Otherwisecluster may not work properly.So if this threshold is exceeded,node will be set as invalid.

Example of 2 node cluster configuration

This section contain example of CloverETL cluster nodes configuration. In addition it is necessssary to configure:

• sharing or replication of directory /home/clover/nfs_shared/sandboxes

• connection to the same database from both nodes

• HTTP load balancer

Configuration of node on 192.168.1.131

jdbc.dialect=org.hibernate.dialect.MySQLDialect datasource.type=JNDI datasource.jndiName=java:comp/env/jdbc/clover_server cluster.enabled=true cluster.node.id=node01 cluster.shared_sandboxes_path=/home/clover/nfs_shared/sandboxes license.file=/home/clover/license/license.dat


127

cluster.group.name=cloverCluster cluster.jgroups.bind_address=192.168.1.131 cluster.jgroups.start_port=7800 cluster.jgroups.tcpping.initial_hosts=192.168.1.13[7800] cluster.http.url=http://192.168.1.131:8080/clover

Configuration of node on 192.168.1.13

jdbc.dialect=org.hibernate.dialect.MySQLDialect datasource.type=JNDI datasource.jndiName=java:comp/env/jdbc/clover_server cluster.enabled=true cluster.node.id=node02 cluster.shared_sandboxes_path=/home/clover/nfs_shared/sandboxes license.file=/home/clover/license/license.dat cluster.group.name=cloverCluster cluster.jgroups.bind_address=192.168.1.13 cluster.jgroups.start_port=7800 cluster.jgroups.tcpping.initial_hosts=192.168.1.131[7800] cluster.http.url=http://192.168.1.13:8080/clover

Load balancing properties

Multiplicators of load balancing criteria. Load balancer decides which cluster node executes graph. It means, thatany node may process request for execution, but graph may be executed on the same or on different node accordingto current load of the nodes and according to these multiplicators.

The higher number, the higher relevance for decision. All multiplicators must be greater then 0.

Each node of the cluster may have different load balancing properties. Any node may process incomming requestsfor transformation execution and each may apply criteria for loadbalancing in a different way according to itsown configuration.

These properties aren't vital for cluster configuration - default values are sufficient


128

Table 24.3. Load balancing properties

property type default description

cluster.lb.balance.running_graphs float 3 Specify importance of running graphs forload balancing.

cluster.lb.balance.memused float 0.5 Specify importance of used memmory forload balancing.

cluster.lb.balance.cpus float 1.5 Specify importance of number of CPUsfor load balancing.

cluster.lb.balance.master_bonus float 1 Specify importance of the fact, that thenode is master. Usually it does notaffect anything, thus value 1 says to loadbalancer: "consider master node the sameas any other node"

cluster.lb.balance.this_node float 2 Specify importance of the fact, thatthe node is the same which processesrequest for execution. The same node,which decides where to execute graph.If you specify this multiplicator greatenough, it will cause, that graph will bealways executed on the same node, whichprocesses request for execution.

129

Chapter 25. Temp Space ManagementMany of the components available in the CloverETL require temporary files or directories in order to workproperly. Temp space is a physical location on local file system where these files or directories are created andmaintained.

Overview

The overview of temp spaces defined in CloverETL Server is available under Configuration > Temp spacemanagement > Overview

Figure 25.1. Configured temp spaces overview - one default temp space on each cluster node

Setup

Temp space management offers an iterface to add, suspend, resume and delete a temp space. It is accessible underConfiguration > Temp space management > Edit.

The screen is divided in two drop-down areas: Global Configuration and Per Node Configuration. The Globalconfiguration manages temp spaces of standalone server or in case of a server cluster temp spaces on all its nodes.The Per Node Configuration allows to maintain temp spaces separately on each node.

Initialization

When CloverETL Server is started the system checks temp space configuration: in case no temp space is configureda new default temp space is created in the directory where java.io.tmpdir system property points. Thedirectory are named as follows:

• ${java.io.tmpdir}/clover_temp in case of a standalone server

Chapter 25. TempSpace Management

130

• ${java.io.tmpdir}/clover_temp_<node_id> in case of server cluster

Adding Temp Space

In order to define new temp space enter its path into text field under lastrow in the table and click the "Add" link.Note that environment variables and system properties may be used in the path, e.g.: ${VARIABLE_NAME}/temp_space If the directory entered does not exist, it will be created.

NoteThe environment variables have higher priority than system properties of the same name. The pathwith varibles are resolved after system has added new temp space and while the server is starting. Incase the variable value has been changed it is necessary to restart the server to such change take effect.

TipThe main point of adding additional temp spaces is to enable higher system throughtput - thereforethe paths entered should point to directories residing on different physical devices to achieve maximalI/O performance.

Figure 25.2. Newly added global temp space using environment property set on both nodes.

Suspending Temp Space

To suspend a temp space click on "Suspend" link in the panel. In case there are files left from previous or currentgraph executions a notification is displayed. Once the temp space has been suspended, no new temporary files willbe created in it, but the files already created may be still used by running jobs.

NoteThe system ensures that at least one active (i.e. not suspended) temp space is available.


131

Figure 25.3. Suspend operation asks for confirmation in case there are data present from running jobs.

Resuming Temp Space

To resume a temp space click on "Resume" link in the panel. Resumed temp space is active, i.e. availabale fortemporal files and directories creation.

Removing Temp Space

To remove a temp space click on "Remove" link in the panel. Only suspended temp space may be removed. Shouldbe there any running jobs using the temp space, system will not allow its removal. In case there are some files leftin the the temp space directory, it is possible to remove them in the displayed notification panel. The availableoptions are:

• Remove - remove temp space from system, but keep its content

• Remove and delete - remove the temp space from system and its content too

• Cancel - do not proceed with operation


132

Figure 25.4. Remove operation asks for confirmation in case there are data present in the temp space.

133

List of Figures2.1. Adjusting Maximum heap size limit ......................................................................................... 112.2. Clover Server as the only running application on IBM Websphere .................................................. 213.1. Sandboxes Section in CloverETL Server Web GUI ..................................................................... 253.2. Sandbox Detail in CloverETL Server Web GUI .......................................................................... 263.3. Sandbox Permissions in CloverETL Server Web GUI .................................................................. 273.4. Web GUI - section "Sandboxes" - context menu on sandbox ......................................................... 283.5. Web GUI - section "Sandboxes" - context menu on folder ............................................................ 283.6. Web GUI - download sandbox in ZIP ....................................................................................... 293.7. Web GUI - upload ZIP to sandbox ........................................................................................... 303.8. Web GUI - upload ZIP results ................................................................................................. 303.9. Web GUI - download file in ZIP ............................................................................................. 313.10. Job config properties ............................................................................................................ 344.1. Executions History - executions table ....................................................................................... 354.2. Executions History - overall perspective .................................................................................... 374.3. Executions Hierarchy with docked list of jobs ........................................................................... 375.1. Web GUI - section "Users" under "Configuration" ..................................................................... 415.2. Web GUI - edit user .............................................................................................................. 425.3. Web GUI - change password .................................................................................................. 425.4. Web GUI - groups assignment ................................................................................................ 425.5. Web GUI - section "Groups" .................................................................................................. 435.6. Web GUI - groups assignment ................................................................................................ 445.7. Tree of permissions ............................................................................................................... 446.1. Web GUI - section "Scheduling" - create new ............................................................................ 456.2. Web GUI - onetime schedule form ........................................................................................... 466.3. Web GUI - schedule form - calendar ........................................................................................ 466.4. Web GUI - periodical schedule form ........................................................................................ 476.5. Cron periodical schedule form ................................................................................................. 486.6. Web GUI - Graph execution task ............................................................................................. 496.7. Web GUI - Jobflow execution task .......................................................................................... 506.8. Web GUI - "Kill job" ............................................................................................................ 516.9. Web GUI - shell command ..................................................................................................... 516.10. Web GUI - archive records ................................................................................................... 547.1. Web GUI - graph timeout event .............................................................................................. 567.2. Web GUI - send email ........................................................................................................... 587.3. Web GUI - Task JMS message editor ....................................................................................... 607.4. Event source graph isn't specified, thus listener works for all graphs in specified sandbox .................... 617.5. Web GUI - email notification about graph failure ....................................................................... 617.6. Web GUI - email notification about graph success ...................................................................... 627.7. Web GUI - backup of data processed by graph ........................................................................... 628.1. Web GUI - jobflow timeout event ............................................................................................ 6411.1. Web GUI - "Manual task execution" section ............................................................................ 7012.1. Web GUI - "File event listeners" section ................................................................................. 7115.1. Glassfish JMX connector ...................................................................................................... 8515.2. Websphere configuration ...................................................................................................... 8615.3. Websphere7 configuration ..................................................................................................... 8717.1. Launch Services and CloverETL Server as web application back-end ............................................ 8917.2. Launch Services section ........................................................................................................ 9117.3. Overview tab ...................................................................................................................... 9117.4. Edit Configuration tab .......................................................................................................... 9217.5. Creating new parameter ........................................................................................................ 9317.6. Edit Parameters tab .............................................................................................................. 9324.1. Dialog form for creating new shared sandbox ......................................................................... 11824.2. Dialog form for creating new local sandbox ........................................................................... 11924.3. List of nodes joined to the cluster ......................................................................................... 12024.4. Cluster Scalability .............................................................................................................. 124

CloverETL Server

134

24.5. Speedup factor .................................................................................................................. 12425.1. Configured temp spaces overview - one default temp space on each cluster node ............................ 12925.2. Newly added global temp space using environment property set on both nodes. ............................. 13025.3. Suspend operation asks for confirmation in case there are data present from running jobs. ................ 13125.4. Remove operation asks for confirmation in case there are data present in the temp space. ................. 132

135

List of Tables1.1. CloverETL Server and CloverETL Engine comparison .................................................................. 23.1. Sandbox attributes ................................................................................................................. 253.2. Sandbox permissions ............................................................................................................. 273.3. ZIP upload parameters ........................................................................................................... 303.4. Job config parameters ............................................................................................................ 334.1. Persistent run record attributes ................................................................................................. 365.1. After default installation above empty DB, there are two users created ............................................ 415.2. User attributes ...................................................................................................................... 415.3. Default groups created during installation .................................................................................. 436.1. Onetime schedule attributes .................................................................................................... 456.2. Periodical schedule attributes .................................................................................................. 476.3. Cron periodical schedule attributes ........................................................................................... 476.4. Attributes of "Graph execution" task ........................................................................................ 496.5. Attributes of "Jobflow execution" task ...................................................................................... 506.6. Attributes of "Kill Job" task .................................................................................................... 516.7. Attributes of "Shell command" task .......................................................................................... 516.8. List of variables available in Groovy code ................................................................................. 526.9. Attributes of "archive records" task .......................................................................................... 537.1. Attributes of "Send email" task ................................................................................................ 577.2. Placeholders useful in email templates ...................................................................................... 597.3. Attributes of JMS message task ............................................................................................... 609.1. Attributes of JMS message task ............................................................................................... 659.2. Variables accessible in groovy code ......................................................................................... 669.3. "properties" elements ............................................................................................................. 679.4. "data" elements ..................................................................................................................... 6710.1. Attributes of Universal message task ....................................................................................... 6910.2. Variables accessible in groovy code ........................................................................................ 6914.1. Parameters of graph_run ....................................................................................................... 7714.2. Parameters of graph_status .................................................................................................... 7814.3. Parameters of graph_kill ....................................................................................................... 7814.4. Parameters of sandbox_content .............................................................................................. 7914.5. Parameters of executions_history ............................................................................................ 8014.6. Parameters of suspend .......................................................................................................... 8114.7. Parameters of resume ........................................................................................................... 8114.8. Parameters of sandbox create ................................................................................................. 8214.9. Parameters of sandbox add location ........................................................................................ 8214.10. Parameters of sandbox add location ...................................................................................... 8318.1. General configuration ......................................................................................................... 10218.2. Defaults for job execution configuration - see section Job config properties for details ..................... 10519.1. Defaults for graph execution configuration - see section Graph config properties for details .............. 10619.2. passed parameters .............................................................................................................. 10719.3. passed parameters .............................................................................................................. 10719.4. passed parameters .............................................................................................................. 10822.1. Variables accessible in groovy code ...................................................................................... 11124.1. Mandatory properties - these properties must be properly set on each node of the cluster .................. 12524.2. Optional properties - these properties aren't vital for cluster configuration - default values aresufficient .................................................................................................................................. 12624.3. Load balancing properties .................................................................................................... 128

oracle® endeca information discovery integrator® endeca information discovery integrator :...

Documents